Skip to content

Instantly share code, notes, and snippets.

Show Gist options
  • Save philipturner/1c157bf87702420e51c6deb68e70d078 to your computer and use it in GitHub Desktop.
Save philipturner/1c157bf87702420e51c6deb68e70d078 to your computer and use it in GitHub Desktop.
High-resolution data for switching from direct (device) to async (threadgroup) stores.
M1 Max, FP32
problemSize = 1488 | A B | 8004 -> 8001 (-3) GFLOPS
problemSize = 1489 | A B | 7538 -> 7585 (+47) GFLOPS
problemSize = 1490 | A B | 7603 -> 7637 (+34) GFLOPS
problemSize = 1491 | A B | 7608 -> 7655 (+47) GFLOPS
problemSize = 1492 | A B | 7679 -> 7700 (+21) GFLOPS
problemSize = 1493 | A B | 7653 -> 7683 (+30) GFLOPS
problemSize = 1494 | A B | 7697 -> 7718 (+21) GFLOPS
problemSize = 1495 | A B | 7700 -> 7730 (+30) GFLOPS
problemSize = 1496 | A B | 7825 -> 7829 (+4) GFLOPS
problemSize = 1497 | A B | 7520 -> 7583 (+63) GFLOPS
problemSize = 1498 | A B | 7580 -> 7618 (+38) GFLOPS
problemSize = 1499 | A B | 7601 -> 7639 (+38) GFLOPS
problemSize = 1500 | A B | 7668 -> 7681 (+13) GFLOPS
problemSize = 1501 | A B | 7651 -> 7672 (+21) GFLOPS
problemSize = 1502 | A B | 7695 -> 7712 (+17) GFLOPS
problemSize = 1503 | A B | 7702 -> 7728 (+26) GFLOPS
problemSize = 1504 | A B | 7793 -> 7804 (+11) GFLOPS
problemSize = 1505 | A B | 7816 -> 7874 (+58) GFLOPS
problemSize = 1506 | A B | 7890 -> 7929 (+39) GFLOPS
problemSize = 1507 | A B | 7907 -> 7960 (+53) GFLOPS
problemSize = 1508 | A B | 7992 -> 8015 (+23) GFLOPS
problemSize = 1509 | A B | 7982 -> 8013 (+31) GFLOPS
problemSize = 1510 | A B | 8025 -> 8047 (+22) GFLOPS
problemSize = 1511 | A B | 8040 -> 8060 (+20) GFLOPS
problemSize = 1512 | A B | 8085 -> 8076 (-9) GFLOPS
problemSize = 1513 | A B | 7681 -> 7734 (+53) GFLOPS
problemSize = 1514 | A B | 7757 -> 7790 (+33) GFLOPS
problemSize = 1515 | A B | 7771 -> 7798 (+27) GFLOPS
problemSize = 1516 | A B | 7834 -> 7857 (+23) GFLOPS
problemSize = 1517 | A B | 7816 -> 7846 (+30) GFLOPS
problemSize = 1518 | A B | 7854 -> 7879 (+25) GFLOPS
problemSize = 1519 | A B | 7850 -> 7881 (+31) GFLOPS
problemSize = 1520 | A B | 7997 -> 8005 (+8) GFLOPS
problemSize = 1521 | A B | 7683 -> 7731 (+48) GFLOPS
problemSize = 1522 | A B | 7726 -> 7774 (+48) GFLOPS
problemSize = 1523 | A B | 7742 -> 7789 (+47) GFLOPS
problemSize = 1524 | A B | 7811 -> 7821 (+10) GFLOPS
problemSize = 1525 | A B | 7798 -> 7827 (+29) GFLOPS
problemSize = 1526 | A B | 7836 -> 7858 (+22) GFLOPS
problemSize = 1527 | A B | 7851 -> 7880 (+29) GFLOPS
problemSize = 1528 | A B | 7948 -> 7944 (-4) GFLOPS
problemSize = 1529 | A B | 7960 -> 8035 (+75) GFLOPS
problemSize = 1530 | A B | 8029 -> 8084 (+55) GFLOPS
problemSize = 1531 | A B | 8043 -> 8103 (+60) GFLOPS
problemSize = 1532 | A B | 8154 -> 8170 (+16) GFLOPS
problemSize = 1533 | A B | 8130 -> 8177 (+47) GFLOPS
problemSize = 1534 | A B | 8192 -> 8205 (+13) GFLOPS
problemSize = 1535 | A B | 8181 -> 8211 (+30) GFLOPS
problemSize = 1536 | A B | 8236 -> 8241 (+5) GFLOPS
problemSize = 1537 | A B | 7557 -> 7614 (+57) GFLOPS
problemSize = 1538 | A B | 7620 -> 7662 (+42) GFLOPS
problemSize = 1539 | A B | 7638 -> 7680 (+42) GFLOPS
problemSize = 1540 | A B | 7711 -> 7727 (+16) GFLOPS
problemSize = 1541 | A B | 7690 -> 7716 (+26) GFLOPS
problemSize = 1542 | A B | 7729 -> 7744 (+15) GFLOPS
problemSize = 1543 | A B | 7733 -> 7759 (+26) GFLOPS
problemSize = 1544 | A B | 7849 -> 7857 (+8) GFLOPS
problemSize = 1545 | A B | 7547 -> 7609 (+62) GFLOPS
problemSize = 1546 | A B | 7601 -> 7643 (+42) GFLOPS
problemSize = 1547 | A B | 7622 -> 7667 (+45) GFLOPS
problemSize = 1548 | A B | 7699 -> 7715 (+16) GFLOPS
problemSize = 1549 | A B | 7691 -> 7718 (+27) GFLOPS
problemSize = 1550 | A B | 7726 -> 7740 (+14) GFLOPS
problemSize = 1551 | A B | 7737 -> 7765 (+28) GFLOPS
problemSize = 1552 | A B | 7830 -> 7825 (-5) GFLOPS
problemSize = 1553 | A B | 7870 -> 7934 (+64) GFLOPS
problemSize = 1554 | A B | 7937 -> 7978 (+41) GFLOPS
problemSize = 1555 | A B | 7952 -> 8007 (+55) GFLOPS
problemSize = 1556 | A B | 8047 -> 8060 (+13) GFLOPS
problemSize = 1557 | A B | 8045 -> 8067 (+22) GFLOPS
problemSize = 1558 | A B | 8085 -> 8094 (+9) GFLOPS
problemSize = 1559 | A B | 8103 -> 8115 (+12) GFLOPS
problemSize = 1560 | A B | 8127 -> 8124 (-3) GFLOPS
problemSize = 1561 | A B | 7608 -> 7656 (+48) GFLOPS
problemSize = 1562 | A B | 7659 -> 7693 (+34) GFLOPS
problemSize = 1563 | A B | 7665 -> 7698 (+33) GFLOPS
problemSize = 1564 | A B | 7730 -> 7747 (+17) GFLOPS
problemSize = 1565 | A B | 7716 -> 7729 (+13) GFLOPS
problemSize = 1566 | A B | 7743 -> 7767 (+24) GFLOPS
problemSize = 1567 | A B | 7746 -> 7780 (+34) GFLOPS
problemSize = 1568 | A B | 7870 -> 7877 (+7) GFLOPS
problemSize = 1569 | A B | 7595 -> 7653 (+58) GFLOPS
problemSize = 1570 | A B | 7650 -> 7692 (+42) GFLOPS
problemSize = 1571 | A B | 7675 -> 7713 (+38) GFLOPS
problemSize = 1572 | A B | 7730 -> 7735 (+5) GFLOPS
problemSize = 1573 | A B | 7720 -> 7748 (+28) GFLOPS
problemSize = 1574 | A B | 7757 -> 7771 (+14) GFLOPS
problemSize = 1575 | A B | 7763 -> 7785 (+22) GFLOPS
problemSize = 1576 | A B | 7833 -> 7837 (+4) GFLOPS
problemSize = 1577 | A B | 7913 -> 7973 (+60) GFLOPS
problemSize = 1578 | A B | 7988 -> 8022 (+34) GFLOPS
problemSize = 1579 | A B | 8000 -> 8046 (+46) GFLOPS
problemSize = 1580 | A B | 8090 -> 8092 (+2) GFLOPS
problemSize = 1581 | A B | 8080 -> 8105 (+25) GFLOPS
problemSize = 1582 | A B | 8127 -> 8130 (+3) GFLOPS
problemSize = 1583 | A B | 8129 -> 8146 (+17) GFLOPS
problemSize = 1584 | A B | 8151 -> 8160 (+9) GFLOPS
problemSize = 1488 | A B^T | 7862 -> 7890 (+28) GFLOPS
problemSize = 1489 | A B^T | 7170 -> 7242 (+72) GFLOPS
problemSize = 1490 | A B^T | 7284 -> 7371 (+87) GFLOPS
problemSize = 1491 | A B^T | 7289 -> 7344 (+55) GFLOPS
problemSize = 1492 | A B^T | 7464 -> 7517 (+53) GFLOPS
problemSize = 1493 | A B^T | 7385 -> 7419 (+34) GFLOPS
problemSize = 1494 | A B^T | 7509 -> 7549 (+40) GFLOPS
problemSize = 1495 | A B^T | 7498 -> 7508 (+10) GFLOPS
problemSize = 1496 | A B^T | 7682 -> 7694 (+12) GFLOPS
problemSize = 1497 | A B^T | 7162 -> 7190 (+28) GFLOPS
problemSize = 1498 | A B^T | 7253 -> 7277 (+24) GFLOPS
problemSize = 1499 | A B^T | 7286 -> 7309 (+23) GFLOPS
problemSize = 1500 | A B^T | 7445 -> 7433 (-12) GFLOPS
problemSize = 1501 | A B^T | 7427 -> 7436 (+9) GFLOPS
problemSize = 1502 | A B^T | 7528 -> 7485 (-43) GFLOPS
problemSize = 1503 | A B^T | 7544 -> 7501 (-43) GFLOPS
problemSize = 1504 | A B^T | 7651 -> 7604 (-47) GFLOPS
problemSize = 1505 | A B^T | 7373 -> 7436 (+63) GFLOPS
problemSize = 1506 | A B^T | 7495 -> 7586 (+91) GFLOPS
problemSize = 1507 | A B^T | 7492 -> 7542 (+50) GFLOPS
problemSize = 1508 | A B^T | 7690 -> 7771 (+81) GFLOPS
problemSize = 1509 | A B^T | 7628 -> 7644 (+16) GFLOPS
problemSize = 1510 | A B^T | 7749 -> 7828 (+79) GFLOPS
problemSize = 1511 | A B^T | 7702 -> 7735 (+33) GFLOPS
problemSize = 1512 | A B^T | 7949 -> 7974 (+25) GFLOPS
problemSize = 1513 | A B^T | 7306 -> 7385 (+79) GFLOPS
problemSize = 1514 | A B^T | 7425 -> 7507 (+82) GFLOPS
problemSize = 1515 | A B^T | 7453 -> 7526 (+73) GFLOPS
problemSize = 1516 | A B^T | 7623 -> 7668 (+45) GFLOPS
problemSize = 1517 | A B^T | 7591 -> 7625 (+34) GFLOPS
problemSize = 1518 | A B^T | 7664 -> 7705 (+41) GFLOPS
problemSize = 1519 | A B^T | 7674 -> 7710 (+36) GFLOPS
problemSize = 1520 | A B^T | 7832 -> 7841 (+9) GFLOPS
problemSize = 1521 | A B^T | 7313 -> 7343 (+30) GFLOPS
problemSize = 1522 | A B^T | 7400 -> 7429 (+29) GFLOPS
problemSize = 1523 | A B^T | 7428 -> 7457 (+29) GFLOPS
problemSize = 1524 | A B^T | 7559 -> 7555 (-4) GFLOPS
problemSize = 1525 | A B^T | 7534 -> 7546 (+12) GFLOPS
problemSize = 1526 | A B^T | 7652 -> 7619 (-33) GFLOPS
problemSize = 1527 | A B^T | 7680 -> 7645 (-35) GFLOPS
problemSize = 1528 | A B^T | 7794 -> 7740 (-54) GFLOPS
problemSize = 1529 | A B^T | 7507 -> 7582 (+75) GFLOPS
problemSize = 1530 | A B^T | 7612 -> 7693 (+81) GFLOPS
problemSize = 1531 | A B^T | 7624 -> 7688 (+64) GFLOPS
problemSize = 1532 | A B^T | 7815 -> 7911 (+96) GFLOPS
problemSize = 1533 | A B^T | 7757 -> 7838 (+81) GFLOPS
problemSize = 1534 | A B^T | 7878 -> 7962 (+84) GFLOPS
problemSize = 1535 | A B^T | 7823 -> 7863 (+40) GFLOPS
problemSize = 1536 | A B^T | 8087 -> 8121 (+34) GFLOPS
problemSize = 1537 | A B^T | 7138 -> 7196 (+58) GFLOPS
problemSize = 1538 | A B^T | 7271 -> 7354 (+83) GFLOPS
problemSize = 1539 | A B^T | 7304 -> 7373 (+69) GFLOPS
problemSize = 1540 | A B^T | 7494 -> 7551 (+57) GFLOPS
problemSize = 1541 | A B^T | 7430 -> 7477 (+47) GFLOPS
problemSize = 1542 | A B^T | 7543 -> 7573 (+30) GFLOPS
problemSize = 1543 | A B^T | 7550 -> 7582 (+32) GFLOPS
problemSize = 1544 | A B^T | 7716 -> 7713 (-3) GFLOPS
problemSize = 1545 | A B^T | 7190 -> 7221 (+31) GFLOPS
problemSize = 1546 | A B^T | 7271 -> 7304 (+33) GFLOPS
problemSize = 1547 | A B^T | 7308 -> 7331 (+23) GFLOPS
problemSize = 1548 | A B^T | 7457 -> 7460 (+3) GFLOPS
problemSize = 1549 | A B^T | 7461 -> 7455 (-6) GFLOPS
problemSize = 1550 | A B^T | 7548 -> 7516 (-32) GFLOPS
problemSize = 1551 | A B^T | 7582 -> 7541 (-41) GFLOPS
problemSize = 1552 | A B^T | 7678 -> 7631 (-47) GFLOPS
problemSize = 1553 | A B^T | 7429 -> 7510 (+81) GFLOPS
problemSize = 1554 | A B^T | 7524 -> 7622 (+98) GFLOPS
problemSize = 1555 | A B^T | 7560 -> 7641 (+81) GFLOPS
problemSize = 1556 | A B^T | 7715 -> 7811 (+96) GFLOPS
problemSize = 1557 | A B^T | 7682 -> 7772 (+90) GFLOPS
problemSize = 1558 | A B^T | 7789 -> 7862 (+73) GFLOPS
problemSize = 1559 | A B^T | 7819 -> 7875 (+56) GFLOPS
problemSize = 1560 | A B^T | 7978 -> 7995 (+17) GFLOPS
problemSize = 1561 | A B^T | 7249 -> 7344 (+95) GFLOPS
problemSize = 1562 | A B^T | 7357 -> 7447 (+90) GFLOPS
problemSize = 1563 | A B^T | 7369 -> 7432 (+63) GFLOPS
problemSize = 1564 | A B^T | 7528 -> 7570 (+42) GFLOPS
problemSize = 1565 | A B^T | 7471 -> 7515 (+44) GFLOPS
problemSize = 1566 | A B^T | 7556 -> 7593 (+37) GFLOPS
problemSize = 1567 | A B^T | 7568 -> 7599 (+31) GFLOPS
problemSize = 1568 | A B^T | 7716 -> 7733 (+17) GFLOPS
problemSize = 1569 | A B^T | 7244 -> 7281 (+37) GFLOPS
problemSize = 1570 | A B^T | 7340 -> 7380 (+40) GFLOPS
problemSize = 1571 | A B^T | 7372 -> 7410 (+38) GFLOPS
problemSize = 1572 | A B^T | 7514 -> 7504 (-10) GFLOPS
problemSize = 1573 | A B^T | 7493 -> 7493 ( 0) GFLOPS
problemSize = 1574 | A B^T | 7581 -> 7547 (-34) GFLOPS
problemSize = 1575 | A B^T | 7611 -> 7566 (-45) GFLOPS
problemSize = 1576 | A B^T | 7698 -> 7646 (-52) GFLOPS
problemSize = 1577 | A B^T | 7472 -> 7548 (+76) GFLOPS
problemSize = 1578 | A B^T | 7572 -> 7670 (+98) GFLOPS
problemSize = 1579 | A B^T | 7603 -> 7694 (+91) GFLOPS
problemSize = 1580 | A B^T | 7759 -> 7843 (+84) GFLOPS
problemSize = 1581 | A B^T | 7726 -> 7822 (+96) GFLOPS
problemSize = 1582 | A B^T | 7832 -> 7907 (+75) GFLOPS
problemSize = 1583 | A B^T | 7860 -> 7927 (+67) GFLOPS
problemSize = 1584 | A B^T | 7997 -> 8009 (+12) GFLOPS
M1 Max, FP16
problemSize = 1488 | A B | 8811 -> 8824 (+13) GFLOPS
problemSize = 1489 | A B | 8439 -> 8460 (+21) GFLOPS
problemSize = 1490 | A B | 8487 -> 8473 (-14) GFLOPS
problemSize = 1491 | A B | 8507 -> 8480 (-27) GFLOPS
problemSize = 1492 | A B | 8525 -> 8529 (+4) GFLOPS
problemSize = 1493 | A B | 8524 -> 8536 (+12) GFLOPS
problemSize = 1494 | A B | 8543 -> 8548 (+5) GFLOPS
problemSize = 1495 | A B | 8554 -> 8567 (+13) GFLOPS
problemSize = 1496 | A B | 8617 -> 8597 (-20) GFLOPS
problemSize = 1497 | A B | 8646 -> 8595 (-51) GFLOPS
problemSize = 1498 | A B | 8643 -> 8594 (-49) GFLOPS
problemSize = 1499 | A B | 8685 -> 8647 (-38) GFLOPS
problemSize = 1500 | A B | 8714 -> 8652 (-62) GFLOPS
problemSize = 1501 | A B | 8698 -> 8663 (-35) GFLOPS
problemSize = 1502 | A B | 8730 -> 8671 (-59) GFLOPS
problemSize = 1503 | A B | 8751 -> 8726 (-25) GFLOPS
problemSize = 1504 | A B | 8814 -> 8775 (-39) GFLOPS
problemSize = 1505 | A B | 8616 -> 8646 (+30) GFLOPS
problemSize = 1506 | A B | 8651 -> 8683 (+32) GFLOPS
problemSize = 1507 | A B | 8693 -> 8677 (-16) GFLOPS
problemSize = 1508 | A B | 8676 -> 8706 (+30) GFLOPS
problemSize = 1509 | A B | 8705 -> 8726 (+21) GFLOPS
problemSize = 1510 | A B | 8724 -> 8745 (+21) GFLOPS
problemSize = 1511 | A B | 8752 -> 8755 (+3) GFLOPS
problemSize = 1512 | A B | 8790 -> 8793 (+3) GFLOPS
problemSize = 1513 | A B | 8637 -> 8632 (-5) GFLOPS
problemSize = 1514 | A B | 8682 -> 8656 (-26) GFLOPS
problemSize = 1515 | A B | 8679 -> 8673 (-6) GFLOPS
problemSize = 1516 | A B | 8712 -> 8704 (-8) GFLOPS
problemSize = 1517 | A B | 8715 -> 8707 (-8) GFLOPS
problemSize = 1518 | A B | 8733 -> 8734 (+1) GFLOPS
problemSize = 1519 | A B | 8750 -> 8755 (+5) GFLOPS
problemSize = 1520 | A B | 8804 -> 8787 (-17) GFLOPS
problemSize = 1521 | A B | 8660 -> 8646 (-14) GFLOPS
problemSize = 1522 | A B | 8697 -> 8677 (-20) GFLOPS
problemSize = 1523 | A B | 8694 -> 8704 (+10) GFLOPS
problemSize = 1524 | A B | 8724 -> 8707 (-17) GFLOPS
problemSize = 1525 | A B | 8737 -> 8734 (-3) GFLOPS
problemSize = 1526 | A B | 8765 -> 8751 (-14) GFLOPS
problemSize = 1527 | A B | 8785 -> 8764 (-21) GFLOPS
problemSize = 1528 | A B | 8810 -> 8800 (-10) GFLOPS
problemSize = 1529 | A B | 8929 -> 8836 (-93) GFLOPS
problemSize = 1530 | A B | 8975 -> 8880 (-95) GFLOPS
problemSize = 1531 | A B | 8991 -> 8885 (-106) GFLOPS
problemSize = 1532 | A B | 9009 -> 8897 (-112) GFLOPS
problemSize = 1533 | A B | 9023 -> 8923 (-100) GFLOPS
problemSize = 1534 | A B | 9046 -> 8967 (-79) GFLOPS
problemSize = 1535 | A B | 9074 -> 8942 (-132) GFLOPS
problemSize = 1536 | A B | 9191 -> 9193 (+2) GFLOPS
problemSize = 1537 | A B | 8542 -> 8561 (+19) GFLOPS
problemSize = 1538 | A B | 8565 -> 8585 (+20) GFLOPS
problemSize = 1539 | A B | 8586 -> 8604 (+18) GFLOPS
problemSize = 1540 | A B | 8604 -> 8628 (+24) GFLOPS
problemSize = 1541 | A B | 8612 -> 8628 (+16) GFLOPS
problemSize = 1542 | A B | 8635 -> 8653 (+18) GFLOPS
problemSize = 1543 | A B | 8649 -> 8679 (+30) GFLOPS
problemSize = 1544 | A B | 8684 -> 8703 (+19) GFLOPS
problemSize = 1545 | A B | 8714 -> 8714 ( 0) GFLOPS
problemSize = 1546 | A B | 8730 -> 8722 (-8) GFLOPS
problemSize = 1547 | A B | 8751 -> 8746 (-5) GFLOPS
problemSize = 1548 | A B | 8778 -> 8771 (-7) GFLOPS
problemSize = 1549 | A B | 8790 -> 8793 (+3) GFLOPS
problemSize = 1550 | A B | 8817 -> 8809 (-8) GFLOPS
problemSize = 1551 | A B | 8833 -> 8816 (-17) GFLOPS
problemSize = 1552 | A B | 8868 -> 8866 (-2) GFLOPS
problemSize = 1553 | A B | 8735 -> 8733 (-2) GFLOPS
problemSize = 1554 | A B | 8758 -> 8756 (-2) GFLOPS
problemSize = 1555 | A B | 8777 -> 8769 (-8) GFLOPS
problemSize = 1556 | A B | 8802 -> 8795 (-7) GFLOPS
problemSize = 1557 | A B | 8823 -> 8813 (-10) GFLOPS
problemSize = 1558 | A B | 8848 -> 8833 (-15) GFLOPS
problemSize = 1559 | A B | 8854 -> 8857 (+3) GFLOPS
problemSize = 1560 | A B | 8879 -> 8868 (-11) GFLOPS
problemSize = 1561 | A B | 8568 -> 8555 (-13) GFLOPS
problemSize = 1562 | A B | 8584 -> 8580 (-4) GFLOPS
problemSize = 1563 | A B | 8605 -> 8599 (-6) GFLOPS
problemSize = 1564 | A B | 8635 -> 8625 (-10) GFLOPS
problemSize = 1565 | A B | 8643 -> 8632 (-11) GFLOPS
problemSize = 1566 | A B | 8665 -> 8654 (-11) GFLOPS
problemSize = 1567 | A B | 8681 -> 8680 (-1) GFLOPS
problemSize = 1568 | A B | 8752 -> 8749 (-3) GFLOPS
problemSize = 1569 | A B | 8693 -> 8671 (-22) GFLOPS
problemSize = 1570 | A B | 8715 -> 8690 (-25) GFLOPS
problemSize = 1571 | A B | 8729 -> 8702 (-27) GFLOPS
problemSize = 1572 | A B | 8758 -> 8726 (-32) GFLOPS
problemSize = 1573 | A B | 8775 -> 8739 (-36) GFLOPS
problemSize = 1574 | A B | 8786 -> 8758 (-28) GFLOPS
problemSize = 1575 | A B | 8802 -> 8775 (-27) GFLOPS
problemSize = 1576 | A B | 8836 -> 8810 (-26) GFLOPS
problemSize = 1577 | A B | 8852 -> 8844 (-8) GFLOPS
problemSize = 1578 | A B | 8880 -> 8866 (-14) GFLOPS
problemSize = 1579 | A B | 8890 -> 8882 (-8) GFLOPS
problemSize = 1580 | A B | 8916 -> 8916 ( 0) GFLOPS
problemSize = 1581 | A B | 8928 -> 8918 (-10) GFLOPS
problemSize = 1582 | A B | 8951 -> 8935 (-16) GFLOPS
problemSize = 1583 | A B | 8973 -> 8954 (-19) GFLOPS
problemSize = 1584 | A B | 8942 -> 8960 (+18) GFLOPS
problemSize = 1488 | A B^T | 8786 -> 8776 (-10) GFLOPS
problemSize = 1489 | A B^T | 8374 -> 8401 (+27) GFLOPS
problemSize = 1490 | A B^T | 8415 -> 8418 (+3) GFLOPS
problemSize = 1491 | A B^T | 8440 -> 8473 (+33) GFLOPS
problemSize = 1492 | A B^T | 8468 -> 8486 (+18) GFLOPS
problemSize = 1493 | A B^T | 8505 -> 8496 (-9) GFLOPS
problemSize = 1494 | A B^T | 8503 -> 8517 (+14) GFLOPS
problemSize = 1495 | A B^T | 8538 -> 8540 (+2) GFLOPS
problemSize = 1496 | A B^T | 8584 -> 8583 (-1) GFLOPS
problemSize = 1497 | A B^T | 8559 -> 8586 (+27) GFLOPS
problemSize = 1498 | A B^T | 8605 -> 8641 (+36) GFLOPS
problemSize = 1499 | A B^T | 8632 -> 8664 (+32) GFLOPS
problemSize = 1500 | A B^T | 8674 -> 8710 (+36) GFLOPS
problemSize = 1501 | A B^T | 8700 -> 8729 (+29) GFLOPS
problemSize = 1502 | A B^T | 8723 -> 8713 (-10) GFLOPS
problemSize = 1503 | A B^T | 8747 -> 8785 (+38) GFLOPS
problemSize = 1504 | A B^T | 8815 -> 8803 (-12) GFLOPS
problemSize = 1505 | A B^T | 8577 -> 8585 (+8) GFLOPS
problemSize = 1506 | A B^T | 8637 -> 8651 (+14) GFLOPS
problemSize = 1507 | A B^T | 8674 -> 8663 (-11) GFLOPS
problemSize = 1508 | A B^T | 8712 -> 8684 (-28) GFLOPS
problemSize = 1509 | A B^T | 8730 -> 8678 (-52) GFLOPS
problemSize = 1510 | A B^T | 8748 -> 8740 (-8) GFLOPS
problemSize = 1511 | A B^T | 8770 -> 8757 (-13) GFLOPS
problemSize = 1512 | A B^T | 8787 -> 8800 (+13) GFLOPS
problemSize = 1513 | A B^T | 8462 -> 8514 (+52) GFLOPS
problemSize = 1514 | A B^T | 8523 -> 8542 (+19) GFLOPS
problemSize = 1515 | A B^T | 8547 -> 8575 (+28) GFLOPS
problemSize = 1516 | A B^T | 8591 -> 8603 (+12) GFLOPS
problemSize = 1517 | A B^T | 8592 -> 8619 (+27) GFLOPS
problemSize = 1518 | A B^T | 8620 -> 8642 (+22) GFLOPS
problemSize = 1519 | A B^T | 8652 -> 8665 (+13) GFLOPS
problemSize = 1520 | A B^T | 8686 -> 8716 (+30) GFLOPS
problemSize = 1521 | A B^T | 8577 -> 8595 (+18) GFLOPS
problemSize = 1522 | A B^T | 8629 -> 8633 (+4) GFLOPS
problemSize = 1523 | A B^T | 8653 -> 8655 (+2) GFLOPS
problemSize = 1524 | A B^T | 8698 -> 8689 (-9) GFLOPS
problemSize = 1525 | A B^T | 8707 -> 8699 (-8) GFLOPS
problemSize = 1526 | A B^T | 8737 -> 8738 (+1) GFLOPS
problemSize = 1527 | A B^T | 8754 -> 8736 (-18) GFLOPS
problemSize = 1528 | A B^T | 8803 -> 8784 (-19) GFLOPS
problemSize = 1529 | A B^T | 8801 -> 8840 (+39) GFLOPS
problemSize = 1530 | A B^T | 8851 -> 8867 (+16) GFLOPS
problemSize = 1531 | A B^T | 8885 -> 8903 (+18) GFLOPS
problemSize = 1532 | A B^T | 8940 -> 8937 (-3) GFLOPS
problemSize = 1533 | A B^T | 8957 -> 8955 (-2) GFLOPS
problemSize = 1534 | A B^T | 8971 -> 8982 (+11) GFLOPS
problemSize = 1535 | A B^T | 8992 -> 9006 (+14) GFLOPS
problemSize = 1536 | A B^T | 9061 -> 9041 (-20) GFLOPS
problemSize = 1537 | A B^T | 8484 -> 8533 (+49) GFLOPS
problemSize = 1538 | A B^T | 8552 -> 8567 (+15) GFLOPS
problemSize = 1539 | A B^T | 8573 -> 8592 (+19) GFLOPS
problemSize = 1540 | A B^T | 8622 -> 8630 (+8) GFLOPS
problemSize = 1541 | A B^T | 8636 -> 8646 (+10) GFLOPS
problemSize = 1542 | A B^T | 8657 -> 8664 (+7) GFLOPS
problemSize = 1543 | A B^T | 8683 -> 8676 (-7) GFLOPS
problemSize = 1544 | A B^T | 8718 -> 8713 (-5) GFLOPS
problemSize = 1545 | A B^T | 8531 -> 8581 (+50) GFLOPS
problemSize = 1546 | A B^T | 8576 -> 8621 (+45) GFLOPS
problemSize = 1547 | A B^T | 8620 -> 8647 (+27) GFLOPS
problemSize = 1548 | A B^T | 8665 -> 8673 (+8) GFLOPS
problemSize = 1549 | A B^T | 8677 -> 8696 (+19) GFLOPS
problemSize = 1550 | A B^T | 8701 -> 8721 (+20) GFLOPS
problemSize = 1551 | A B^T | 8718 -> 8734 (+16) GFLOPS
problemSize = 1552 | A B^T | 8783 -> 8794 (+11) GFLOPS
problemSize = 1553 | A B^T | 8655 -> 8675 (+20) GFLOPS
problemSize = 1554 | A B^T | 8701 -> 8706 (+5) GFLOPS
problemSize = 1555 | A B^T | 8735 -> 8738 (+3) GFLOPS
problemSize = 1556 | A B^T | 8774 -> 8773 (-1) GFLOPS
problemSize = 1557 | A B^T | 8785 -> 8781 (-4) GFLOPS
problemSize = 1558 | A B^T | 8821 -> 8810 (-11) GFLOPS
problemSize = 1559 | A B^T | 8830 -> 8834 (+4) GFLOPS
problemSize = 1560 | A B^T | 8885 -> 8820 (-65) GFLOPS
problemSize = 1561 | A B^T | 8704 -> 8733 (+29) GFLOPS
problemSize = 1562 | A B^T | 8759 -> 8775 (+16) GFLOPS
problemSize = 1563 | A B^T | 8782 -> 8798 (+16) GFLOPS
problemSize = 1564 | A B^T | 8821 -> 8829 (+8) GFLOPS
problemSize = 1565 | A B^T | 8846 -> 8848 (+2) GFLOPS
problemSize = 1566 | A B^T | 8861 -> 8877 (+16) GFLOPS
problemSize = 1567 | A B^T | 8876 -> 8888 (+12) GFLOPS
problemSize = 1568 | A B^T | 8951 -> 8956 (+5) GFLOPS
problemSize = 1569 | A B^T | 8681 -> 8712 (+31) GFLOPS
problemSize = 1570 | A B^T | 8714 -> 8752 (+38) GFLOPS
problemSize = 1571 | A B^T | 8751 -> 8767 (+16) GFLOPS
problemSize = 1572 | A B^T | 8793 -> 8795 (+2) GFLOPS
problemSize = 1573 | A B^T | 8802 -> 8806 (+4) GFLOPS
problemSize = 1574 | A B^T | 8822 -> 8833 (+11) GFLOPS
problemSize = 1575 | A B^T | 8846 -> 8843 (-3) GFLOPS
problemSize = 1576 | A B^T | 8883 -> 8881 (-2) GFLOPS
problemSize = 1577 | A B^T | 8685 -> 8733 (+48) GFLOPS
problemSize = 1578 | A B^T | 8729 -> 8765 (+36) GFLOPS
problemSize = 1579 | A B^T | 8756 -> 8797 (+41) GFLOPS
problemSize = 1580 | A B^T | 8793 -> 8824 (+31) GFLOPS
problemSize = 1581 | A B^T | 8811 -> 8842 (+31) GFLOPS
problemSize = 1582 | A B^T | 8840 -> 8861 (+21) GFLOPS
problemSize = 1583 | A B^T | 8850 -> 8887 (+37) GFLOPS
problemSize = 1584 | A B^T | 8927 -> 8929 (+2) GFLOPS
M1 Max, BF16
problemSize = 1488 | A B | 8039 -> 8066 (+27) GFLOPS
problemSize = 1489 | A B | 7558 -> 7549 (-9) GFLOPS
problemSize = 1490 | A B | 7567 -> 7572 (+5) GFLOPS
problemSize = 1491 | A B | 7599 -> 7590 (-9) GFLOPS
problemSize = 1492 | A B | 7608 -> 7610 (+2) GFLOPS
problemSize = 1493 | A B | 7631 -> 7631 ( 0) GFLOPS
problemSize = 1494 | A B | 7637 -> 7649 (+12) GFLOPS
problemSize = 1495 | A B | 7665 -> 7666 (+1) GFLOPS
problemSize = 1496 | A B | 7696 -> 7701 (+5) GFLOPS
problemSize = 1497 | A B | 7852 -> 7628 (-224) GFLOPS
problemSize = 1498 | A B | 7893 -> 7655 (-238) GFLOPS
problemSize = 1499 | A B | 7888 -> 7671 (-217) GFLOPS
problemSize = 1500 | A B | 7939 -> 7705 (-234) GFLOPS
problemSize = 1501 | A B | 7933 -> 7707 (-226) GFLOPS
problemSize = 1502 | A B | 7973 -> 7729 (-244) GFLOPS
problemSize = 1503 | A B | 7970 -> 7750 (-220) GFLOPS
problemSize = 1504 | A B | 8067 -> 7834 (-233) GFLOPS
problemSize = 1505 | A B | 7840 -> 7864 (+24) GFLOPS
problemSize = 1506 | A B | 7883 -> 7882 (-1) GFLOPS
problemSize = 1507 | A B | 7878 -> 7892 (+14) GFLOPS
problemSize = 1508 | A B | 7935 -> 7925 (-10) GFLOPS
problemSize = 1509 | A B | 7915 -> 7933 (+18) GFLOPS
problemSize = 1510 | A B | 7959 -> 7941 (-18) GFLOPS
problemSize = 1511 | A B | 7940 -> 7968 (+28) GFLOPS
problemSize = 1512 | A B | 7995 -> 8003 (+8) GFLOPS
problemSize = 1513 | A B | 7843 -> 7787 (-56) GFLOPS
problemSize = 1514 | A B | 7916 -> 7815 (-101) GFLOPS
problemSize = 1515 | A B | 7875 -> 7837 (-38) GFLOPS
problemSize = 1516 | A B | 7945 -> 7854 (-91) GFLOPS
problemSize = 1517 | A B | 7912 -> 7863 (-49) GFLOPS
problemSize = 1518 | A B | 8002 -> 7890 (-112) GFLOPS
problemSize = 1519 | A B | 7944 -> 7895 (-49) GFLOPS
problemSize = 1520 | A B | 8054 -> 7963 (-91) GFLOPS
problemSize = 1521 | A B | 7796 -> 7791 (-5) GFLOPS
problemSize = 1522 | A B | 7817 -> 7809 (-8) GFLOPS
problemSize = 1523 | A B | 7835 -> 7831 (-4) GFLOPS
problemSize = 1524 | A B | 7853 -> 7854 (+1) GFLOPS
problemSize = 1525 | A B | 7866 -> 7858 (-8) GFLOPS
problemSize = 1526 | A B | 7878 -> 7894 (+16) GFLOPS
problemSize = 1527 | A B | 7914 -> 7898 (-16) GFLOPS
problemSize = 1528 | A B | 7941 -> 7944 (+3) GFLOPS
problemSize = 1529 | A B | 8099 -> 7862 (-237) GFLOPS
problemSize = 1530 | A B | 8149 -> 7882 (-267) GFLOPS
problemSize = 1531 | A B | 8139 -> 7896 (-243) GFLOPS
problemSize = 1532 | A B | 8188 -> 7938 (-250) GFLOPS
problemSize = 1533 | A B | 8187 -> 7937 (-250) GFLOPS
problemSize = 1534 | A B | 8224 -> 7968 (-256) GFLOPS
problemSize = 1535 | A B | 8219 -> 7980 (-239) GFLOPS
problemSize = 1536 | A B | 8351 -> 8347 (-4) GFLOPS
problemSize = 1537 | A B | 7693 -> 7709 (+16) GFLOPS
problemSize = 1538 | A B | 7744 -> 7729 (-15) GFLOPS
problemSize = 1539 | A B | 7741 -> 7738 (-3) GFLOPS
problemSize = 1540 | A B | 7787 -> 7771 (-16) GFLOPS
problemSize = 1541 | A B | 7778 -> 7782 (+4) GFLOPS
problemSize = 1542 | A B | 7820 -> 7786 (-34) GFLOPS
problemSize = 1543 | A B | 7810 -> 7816 (+6) GFLOPS
problemSize = 1544 | A B | 7874 -> 7842 (-32) GFLOPS
problemSize = 1545 | A B | 7952 -> 7742 (-210) GFLOPS
problemSize = 1546 | A B | 7979 -> 7765 (-214) GFLOPS
problemSize = 1547 | A B | 7998 -> 7776 (-222) GFLOPS
problemSize = 1548 | A B | 8013 -> 7797 (-216) GFLOPS
problemSize = 1549 | A B | 8028 -> 7812 (-216) GFLOPS
problemSize = 1550 | A B | 8054 -> 7828 (-226) GFLOPS
problemSize = 1551 | A B | 8064 -> 7853 (-211) GFLOPS
problemSize = 1552 | A B | 8113 -> 7905 (-208) GFLOPS
problemSize = 1553 | A B | 7763 -> 7737 (-26) GFLOPS
problemSize = 1554 | A B | 7767 -> 7765 (-2) GFLOPS
problemSize = 1555 | A B | 7796 -> 7774 (-22) GFLOPS
problemSize = 1556 | A B | 7812 -> 7793 (-19) GFLOPS
problemSize = 1557 | A B | 7833 -> 7809 (-24) GFLOPS
problemSize = 1558 | A B | 7841 -> 7828 (-13) GFLOPS
problemSize = 1559 | A B | 7879 -> 7846 (-33) GFLOPS
problemSize = 1560 | A B | 7887 -> 7874 (-13) GFLOPS
problemSize = 1561 | A B | 7826 -> 7613 (-213) GFLOPS
problemSize = 1562 | A B | 7870 -> 7636 (-234) GFLOPS
problemSize = 1563 | A B | 7860 -> 7651 (-209) GFLOPS
problemSize = 1564 | A B | 7907 -> 7686 (-221) GFLOPS
problemSize = 1565 | A B | 7895 -> 7681 (-214) GFLOPS
problemSize = 1566 | A B | 7939 -> 7698 (-241) GFLOPS
problemSize = 1567 | A B | 7932 -> 7715 (-217) GFLOPS
problemSize = 1568 | A B | 8021 -> 7791 (-230) GFLOPS
problemSize = 1569 | A B | 7799 -> 7828 (+29) GFLOPS
problemSize = 1570 | A B | 7854 -> 7839 (-15) GFLOPS
problemSize = 1571 | A B | 7835 -> 7856 (+21) GFLOPS
problemSize = 1572 | A B | 7886 -> 7888 (+2) GFLOPS
problemSize = 1573 | A B | 7873 -> 7889 (+16) GFLOPS
problemSize = 1574 | A B | 7919 -> 7908 (-11) GFLOPS
problemSize = 1575 | A B | 7914 -> 7929 (+15) GFLOPS
problemSize = 1576 | A B | 7975 -> 7965 (-10) GFLOPS
problemSize = 1577 | A B | 8041 -> 7851 (-190) GFLOPS
problemSize = 1578 | A B | 8067 -> 7879 (-188) GFLOPS
problemSize = 1579 | A B | 8085 -> 7893 (-192) GFLOPS
problemSize = 1580 | A B | 8094 -> 7911 (-183) GFLOPS
problemSize = 1581 | A B | 8106 -> 7920 (-186) GFLOPS
problemSize = 1582 | A B | 8123 -> 7943 (-180) GFLOPS
problemSize = 1583 | A B | 8137 -> 7953 (-184) GFLOPS
problemSize = 1584 | A B | 8149 -> 8171 (+22) GFLOPS
problemSize = 1488 | A B^T | 8287 -> 8286 (-1) GFLOPS
problemSize = 1489 | A B^T | 7827 -> 7794 (-33) GFLOPS
problemSize = 1490 | A B^T | 7794 -> 7828 (+34) GFLOPS
problemSize = 1491 | A B^T | 7903 -> 7857 (-46) GFLOPS
problemSize = 1492 | A B^T | 7854 -> 7885 (+31) GFLOPS
problemSize = 1493 | A B^T | 7926 -> 7898 (-28) GFLOPS
problemSize = 1494 | A B^T | 7897 -> 7928 (+31) GFLOPS
problemSize = 1495 | A B^T | 7987 -> 7952 (-35) GFLOPS
problemSize = 1496 | A B^T | 7958 -> 7985 (+27) GFLOPS
problemSize = 1497 | A B^T | 7835 -> 7562 (-273) GFLOPS
problemSize = 1498 | A B^T | 7909 -> 7603 (-306) GFLOPS
problemSize = 1499 | A B^T | 7902 -> 7638 (-264) GFLOPS
problemSize = 1500 | A B^T | 7975 -> 7671 (-304) GFLOPS
problemSize = 1501 | A B^T | 7955 -> 7691 (-264) GFLOPS
problemSize = 1502 | A B^T | 8001 -> 7716 (-285) GFLOPS
problemSize = 1503 | A B^T | 7989 -> 7731 (-258) GFLOPS
problemSize = 1504 | A B^T | 8148 -> 7864 (-284) GFLOPS
problemSize = 1505 | A B^T | 7879 -> 8033 (+154) GFLOPS
problemSize = 1506 | A B^T | 8030 -> 8067 (+37) GFLOPS
problemSize = 1507 | A B^T | 7939 -> 8090 (+151) GFLOPS
problemSize = 1508 | A B^T | 8093 -> 8119 (+26) GFLOPS
problemSize = 1509 | A B^T | 7985 -> 8143 (+158) GFLOPS
problemSize = 1510 | A B^T | 8125 -> 8151 (+26) GFLOPS
problemSize = 1511 | A B^T | 8020 -> 8169 (+149) GFLOPS
problemSize = 1512 | A B^T | 8145 -> 8209 (+64) GFLOPS
problemSize = 1513 | A B^T | 8089 -> 7918 (-171) GFLOPS
problemSize = 1514 | A B^T | 8173 -> 7963 (-210) GFLOPS
problemSize = 1515 | A B^T | 8191 -> 7979 (-212) GFLOPS
problemSize = 1516 | A B^T | 8260 -> 8026 (-234) GFLOPS
problemSize = 1517 | A B^T | 8230 -> 8041 (-189) GFLOPS
problemSize = 1518 | A B^T | 8278 -> 8063 (-215) GFLOPS
problemSize = 1519 | A B^T | 8280 -> 8074 (-206) GFLOPS
problemSize = 1520 | A B^T | 8359 -> 8140 (-219) GFLOPS
problemSize = 1521 | A B^T | 8037 -> 8032 (-5) GFLOPS
problemSize = 1522 | A B^T | 8039 -> 8063 (+24) GFLOPS
problemSize = 1523 | A B^T | 8064 -> 8087 (+23) GFLOPS
problemSize = 1524 | A B^T | 8097 -> 8126 (+29) GFLOPS
problemSize = 1525 | A B^T | 8113 -> 8137 (+24) GFLOPS
problemSize = 1526 | A B^T | 8133 -> 8163 (+30) GFLOPS
problemSize = 1527 | A B^T | 8179 -> 8183 (+4) GFLOPS
problemSize = 1528 | A B^T | 8211 -> 8242 (+31) GFLOPS
problemSize = 1529 | A B^T | 8060 -> 7766 (-294) GFLOPS
problemSize = 1530 | A B^T | 8130 -> 7818 (-312) GFLOPS
problemSize = 1531 | A B^T | 8134 -> 7851 (-283) GFLOPS
problemSize = 1532 | A B^T | 8206 -> 7888 (-318) GFLOPS
problemSize = 1533 | A B^T | 8190 -> 7909 (-281) GFLOPS
problemSize = 1534 | A B^T | 8252 -> 7934 (-318) GFLOPS
problemSize = 1535 | A B^T | 8243 -> 7950 (-293) GFLOPS
problemSize = 1536 | A B^T | 8398 -> 8397 (-1) GFLOPS
problemSize = 1537 | A B^T | 7753 -> 7905 (+152) GFLOPS
problemSize = 1538 | A B^T | 7897 -> 7946 (+49) GFLOPS
problemSize = 1539 | A B^T | 7810 -> 7964 (+154) GFLOPS
problemSize = 1540 | A B^T | 7959 -> 7984 (+25) GFLOPS
problemSize = 1541 | A B^T | 7845 -> 8003 (+158) GFLOPS
problemSize = 1542 | A B^T | 7993 -> 8023 (+30) GFLOPS
problemSize = 1543 | A B^T | 7897 -> 8037 (+140) GFLOPS
problemSize = 1544 | A B^T | 8054 -> 8082 (+28) GFLOPS
problemSize = 1545 | A B^T | 7904 -> 7941 (+37) GFLOPS
problemSize = 1546 | A B^T | 7994 -> 7970 (-24) GFLOPS
problemSize = 1547 | A B^T | 7962 -> 7986 (+24) GFLOPS
problemSize = 1548 | A B^T | 8045 -> 8017 (-28) GFLOPS
problemSize = 1549 | A B^T | 8016 -> 8033 (+17) GFLOPS
problemSize = 1550 | A B^T | 8084 -> 8047 (-37) GFLOPS
problemSize = 1551 | A B^T | 8052 -> 8077 (+25) GFLOPS
problemSize = 1552 | A B^T | 8137 -> 8138 (+1) GFLOPS
problemSize = 1553 | A B^T | 8097 -> 7995 (-102) GFLOPS
problemSize = 1554 | A B^T | 8013 -> 8034 (+21) GFLOPS
problemSize = 1555 | A B^T | 8161 -> 8057 (-104) GFLOPS
problemSize = 1556 | A B^T | 8066 -> 8096 (+30) GFLOPS
problemSize = 1557 | A B^T | 8209 -> 8107 (-102) GFLOPS
problemSize = 1558 | A B^T | 8101 -> 8128 (+27) GFLOPS
problemSize = 1559 | A B^T | 8267 -> 8146 (-121) GFLOPS
problemSize = 1560 | A B^T | 8175 -> 8124 (-51) GFLOPS
problemSize = 1561 | A B^T | 7819 -> 7551 (-268) GFLOPS
problemSize = 1562 | A B^T | 7884 -> 7596 (-288) GFLOPS
problemSize = 1563 | A B^T | 7875 -> 7611 (-264) GFLOPS
problemSize = 1564 | A B^T | 7943 -> 7651 (-292) GFLOPS
problemSize = 1565 | A B^T | 7916 -> 7663 (-253) GFLOPS
problemSize = 1566 | A B^T | 7982 -> 7688 (-294) GFLOPS
problemSize = 1567 | A B^T | 7959 -> 7707 (-252) GFLOPS
problemSize = 1568 | A B^T | 8105 -> 7816 (-289) GFLOPS
problemSize = 1569 | A B^T | 7852 -> 8004 (+152) GFLOPS
problemSize = 1570 | A B^T | 8007 -> 8039 (+32) GFLOPS
problemSize = 1571 | A B^T | 7908 -> 8056 (+148) GFLOPS
problemSize = 1572 | A B^T | 8056 -> 8084 (+28) GFLOPS
problemSize = 1573 | A B^T | 7943 -> 8099 (+156) GFLOPS
problemSize = 1574 | A B^T | 8089 -> 8117 (+28) GFLOPS
problemSize = 1575 | A B^T | 7986 -> 8132 (+146) GFLOPS
problemSize = 1576 | A B^T | 8152 -> 8177 (+25) GFLOPS
problemSize = 1577 | A B^T | 8068 -> 8023 (-45) GFLOPS
problemSize = 1578 | A B^T | 8129 -> 8052 (-77) GFLOPS
problemSize = 1579 | A B^T | 8106 -> 8076 (-30) GFLOPS
problemSize = 1580 | A B^T | 8180 -> 8098 (-82) GFLOPS
problemSize = 1581 | A B^T | 8153 -> 8114 (-39) GFLOPS
problemSize = 1582 | A B^T | 8215 -> 8137 (-78) GFLOPS
problemSize = 1583 | A B^T | 8186 -> 8152 (-34) GFLOPS
problemSize = 1584 | A B^T | 8284 -> 8277 (-7) GFLOPS
M1 Max, FP32
problemSize = 976 | A B | 6925 -> 6920 (-5) GFLOPS
problemSize = 977 | A B | 6851 -> 6883 (+32) GFLOPS
problemSize = 978 | A B | 6908 -> 6938 (+30) GFLOPS
problemSize = 979 | A B | 6955 -> 6982 (+27) GFLOPS
problemSize = 980 | A B | 7026 -> 7007 (-19) GFLOPS
problemSize = 981 | A B | 7012 -> 7028 (+16) GFLOPS
problemSize = 982 | A B | 7059 -> 7106 (+47) GFLOPS
problemSize = 983 | A B | 7083 -> 7083 ( 0) GFLOPS
problemSize = 984 | A B | 7106 -> 7098 (-8) GFLOPS
problemSize = 985 | A B | 6702 -> 6747 (+45) GFLOPS
problemSize = 986 | A B | 6756 -> 6847 (+91) GFLOPS
problemSize = 987 | A B | 6757 -> 6814 (+57) GFLOPS
problemSize = 988 | A B | 6862 -> 6915 (+53) GFLOPS
problemSize = 989 | A B | 6826 -> 6884 (+58) GFLOPS
problemSize = 990 | A B | 6887 -> 6928 (+41) GFLOPS
problemSize = 991 | A B | 6878 -> 6964 (+86) GFLOPS
problemSize = 992 | A B | 7084 -> 7124 (+40) GFLOPS
problemSize = 993 | A B | 6743 -> 6776 (+33) GFLOPS
problemSize = 994 | A B | 6826 -> 6873 (+47) GFLOPS
problemSize = 995 | A B | 6850 -> 6881 (+31) GFLOPS
problemSize = 996 | A B | 6955 -> 6986 (+31) GFLOPS
problemSize = 997 | A B | 6900 -> 6936 (+36) GFLOPS
problemSize = 998 | A B | 6949 -> 7068 (+119) GFLOPS
problemSize = 999 | A B | 6966 -> 6995 (+29) GFLOPS
problemSize = 1000 | A B | 7100 -> 7095 (-5) GFLOPS
problemSize = 1001 | A B | 6938 -> 6965 (+27) GFLOPS
problemSize = 1002 | A B | 7024 -> 7047 (+23) GFLOPS
problemSize = 1003 | A B | 7035 -> 7069 (+34) GFLOPS
problemSize = 1004 | A B | 7146 -> 7173 (+27) GFLOPS
problemSize = 1005 | A B | 7111 -> 7123 (+12) GFLOPS
problemSize = 1006 | A B | 7171 -> 7188 (+17) GFLOPS
problemSize = 1007 | A B | 7188 -> 7209 (+21) GFLOPS
problemSize = 1008 | A B | 7235 -> 7251 (+16) GFLOPS
problemSize = 1009 | A B | 6608 -> 6671 (+63) GFLOPS
problemSize = 1010 | A B | 6696 -> 6719 (+23) GFLOPS
problemSize = 1011 | A B | 6702 -> 6749 (+47) GFLOPS
problemSize = 1012 | A B | 6778 -> 6807 (+29) GFLOPS
problemSize = 1013 | A B | 6800 -> 6812 (+12) GFLOPS
problemSize = 1014 | A B | 6831 -> 6852 (+21) GFLOPS
problemSize = 1015 | A B | 6846 -> 6861 (+15) GFLOPS
problemSize = 1016 | A B | 6957 -> 6964 (+7) GFLOPS
problemSize = 1017 | A B | 6684 -> 6734 (+50) GFLOPS
problemSize = 1018 | A B | 6738 -> 6775 (+37) GFLOPS
problemSize = 1019 | A B | 6773 -> 6805 (+32) GFLOPS
problemSize = 1020 | A B | 6842 -> 6851 (+9) GFLOPS
problemSize = 1021 | A B | 6837 -> 6865 (+28) GFLOPS
problemSize = 1022 | A B | 6875 -> 6904 (+29) GFLOPS
problemSize = 1023 | A B | 6890 -> 6912 (+22) GFLOPS
problemSize = 1024 | A B | 6994 -> 6995 (+1) GFLOPS
problemSize = 1025 | A B | 7023 -> 7097 (+74) GFLOPS
problemSize = 1026 | A B | 7080 -> 7123 (+43) GFLOPS
problemSize = 1027 | A B | 7121 -> 7175 (+54) GFLOPS
problemSize = 1028 | A B | 7202 -> 7212 (+10) GFLOPS
problemSize = 1029 | A B | 7189 -> 7223 (+34) GFLOPS
problemSize = 1030 | A B | 7237 -> 7280 (+43) GFLOPS
problemSize = 1031 | A B | 7264 -> 7292 (+28) GFLOPS
problemSize = 1032 | A B | 7312 -> 7320 (+8) GFLOPS
problemSize = 1033 | A B | 6870 -> 6933 (+63) GFLOPS
problemSize = 1034 | A B | 6939 -> 6975 (+36) GFLOPS
problemSize = 1035 | A B | 6951 -> 7027 (+76) GFLOPS
problemSize = 1036 | A B | 7045 -> 7056 (+11) GFLOPS
problemSize = 1037 | A B | 7016 -> 7049 (+33) GFLOPS
problemSize = 1038 | A B | 7063 -> 7085 (+22) GFLOPS
problemSize = 1039 | A B | 7060 -> 7087 (+27) GFLOPS
problemSize = 1040 | A B | 7193 -> 7194 (+1) GFLOPS
problemSize = 1041 | A B | 6867 -> 6933 (+66) GFLOPS
problemSize = 1042 | A B | 6954 -> 6993 (+39) GFLOPS
problemSize = 1043 | A B | 7004 -> 7047 (+43) GFLOPS
problemSize = 1044 | A B | 7040 -> 7113 (+73) GFLOPS
problemSize = 1045 | A B | 7057 -> 7095 (+38) GFLOPS
problemSize = 1046 | A B | 7101 -> 7106 (+5) GFLOPS
problemSize = 1047 | A B | 7129 -> 7137 (+8) GFLOPS
problemSize = 1048 | A B | 7221 -> 7217 (-4) GFLOPS
problemSize = 1049 | A B | 7232 -> 7280 (+48) GFLOPS
problemSize = 1050 | A B | 7306 -> 7345 (+39) GFLOPS
problemSize = 1051 | A B | 7341 -> 7370 (+29) GFLOPS
problemSize = 1052 | A B | 7413 -> 7450 (+37) GFLOPS
problemSize = 1053 | A B | 7431 -> 7457 (+26) GFLOPS
problemSize = 1054 | A B | 7494 -> 7496 (+2) GFLOPS
problemSize = 1055 | A B | 7524 -> 7493 (-31) GFLOPS
problemSize = 1056 | A B | 7501 -> 7525 (+24) GFLOPS
problemSize = 1057 | A B | 6920 -> 7001 (+81) GFLOPS
problemSize = 1058 | A B | 7039 -> 7040 (+1) GFLOPS
problemSize = 1059 | A B | 7035 -> 7070 (+35) GFLOPS
problemSize = 1060 | A B | 7102 -> 7124 (+22) GFLOPS
problemSize = 1061 | A B | 7101 -> 7119 (+18) GFLOPS
problemSize = 1062 | A B | 7150 -> 7163 (+13) GFLOPS
problemSize = 1063 | A B | 7144 -> 7163 (+19) GFLOPS
problemSize = 1064 | A B | 7270 -> 7270 ( 0) GFLOPS
problemSize = 1065 | A B | 6957 -> 7021 (+64) GFLOPS
problemSize = 1066 | A B | 7020 -> 7052 (+32) GFLOPS
problemSize = 1067 | A B | 7052 -> 7076 (+24) GFLOPS
problemSize = 1068 | A B | 7133 -> 7144 (+11) GFLOPS
problemSize = 1069 | A B | 7116 -> 7138 (+22) GFLOPS
problemSize = 1070 | A B | 7153 -> 7176 (+23) GFLOPS
problemSize = 1071 | A B | 7168 -> 7196 (+28) GFLOPS
problemSize = 1072 | A B | 7276 -> 7284 (+8) GFLOPS
problemSize = 976 | A B^T | 6752 -> 6719 (-33) GFLOPS
problemSize = 977 | A B^T | 6440 -> 6515 (+75) GFLOPS
problemSize = 978 | A B^T | 6542 -> 6634 (+92) GFLOPS
problemSize = 979 | A B^T | 6581 -> 6685 (+104) GFLOPS
problemSize = 980 | A B^T | 6721 -> 6795 (+74) GFLOPS
problemSize = 981 | A B^T | 6744 -> 6804 (+60) GFLOPS
problemSize = 982 | A B^T | 6836 -> 6902 (+66) GFLOPS
problemSize = 983 | A B^T | 6862 -> 6926 (+64) GFLOPS
problemSize = 984 | A B^T | 7042 -> 7091 (+49) GFLOPS
problemSize = 985 | A B^T | 6320 -> 6383 (+63) GFLOPS
problemSize = 986 | A B^T | 6398 -> 6534 (+136) GFLOPS
problemSize = 987 | A B^T | 6453 -> 6504 (+51) GFLOPS
problemSize = 988 | A B^T | 6641 -> 6692 (+51) GFLOPS
problemSize = 989 | A B^T | 6610 -> 6658 (+48) GFLOPS
problemSize = 990 | A B^T | 6700 -> 6747 (+47) GFLOPS
problemSize = 991 | A B^T | 6680 -> 6782 (+102) GFLOPS
problemSize = 992 | A B^T | 6920 -> 7011 (+91) GFLOPS
problemSize = 993 | A B^T | 6318 -> 6372 (+54) GFLOPS
problemSize = 994 | A B^T | 6448 -> 6495 (+47) GFLOPS
problemSize = 995 | A B^T | 6509 -> 6509 ( 0) GFLOPS
problemSize = 996 | A B^T | 6676 -> 6655 (-21) GFLOPS
problemSize = 997 | A B^T | 6628 -> 6654 (+26) GFLOPS
problemSize = 998 | A B^T | 6745 -> 6745 ( 0) GFLOPS
problemSize = 999 | A B^T | 6754 -> 6736 (-18) GFLOPS
problemSize = 1000 | A B^T | 6925 -> 6901 (-24) GFLOPS
problemSize = 1001 | A B^T | 6542 -> 6651 (+109) GFLOPS
problemSize = 1002 | A B^T | 6659 -> 6735 (+76) GFLOPS
problemSize = 1003 | A B^T | 6682 -> 6792 (+110) GFLOPS
problemSize = 1004 | A B^T | 6850 -> 6930 (+80) GFLOPS
problemSize = 1005 | A B^T | 6808 -> 6877 (+69) GFLOPS
problemSize = 1006 | A B^T | 6925 -> 6989 (+64) GFLOPS
problemSize = 1007 | A B^T | 6987 -> 7035 (+48) GFLOPS
problemSize = 1008 | A B^T | 7131 -> 7171 (+40) GFLOPS
problemSize = 1009 | A B^T | 6262 -> 6332 (+70) GFLOPS
problemSize = 1010 | A B^T | 6375 -> 6447 (+72) GFLOPS
problemSize = 1011 | A B^T | 6413 -> 6477 (+64) GFLOPS
problemSize = 1012 | A B^T | 6596 -> 6637 (+41) GFLOPS
problemSize = 1013 | A B^T | 6595 -> 6619 (+24) GFLOPS
problemSize = 1014 | A B^T | 6669 -> 6708 (+39) GFLOPS
problemSize = 1015 | A B^T | 6692 -> 6719 (+27) GFLOPS
problemSize = 1016 | A B^T | 6836 -> 6850 (+14) GFLOPS
problemSize = 1017 | A B^T | 6350 -> 6375 (+25) GFLOPS
problemSize = 1018 | A B^T | 6430 -> 6460 (+30) GFLOPS
problemSize = 1019 | A B^T | 6486 -> 6503 (+17) GFLOPS
problemSize = 1020 | A B^T | 6631 -> 6620 (-11) GFLOPS
problemSize = 1021 | A B^T | 6617 -> 6625 (+8) GFLOPS
problemSize = 1022 | A B^T | 6698 -> 6682 (-16) GFLOPS
problemSize = 1023 | A B^T | 6743 -> 6734 (-9) GFLOPS
problemSize = 1024 | A B^T | 6853 -> 6820 (-33) GFLOPS
problemSize = 1025 | A B^T | 6633 -> 6689 (+56) GFLOPS
problemSize = 1026 | A B^T | 6700 -> 6778 (+78) GFLOPS
problemSize = 1027 | A B^T | 6758 -> 6833 (+75) GFLOPS
problemSize = 1028 | A B^T | 6909 -> 6971 (+62) GFLOPS
problemSize = 1029 | A B^T | 6894 -> 6964 (+70) GFLOPS
problemSize = 1030 | A B^T | 7007 -> 7048 (+41) GFLOPS
problemSize = 1031 | A B^T | 7027 -> 7080 (+53) GFLOPS
problemSize = 1032 | A B^T | 7200 -> 7209 (+9) GFLOPS
problemSize = 1033 | A B^T | 6519 -> 6604 (+85) GFLOPS
problemSize = 1034 | A B^T | 6625 -> 6704 (+79) GFLOPS
problemSize = 1035 | A B^T | 6655 -> 6739 (+84) GFLOPS
problemSize = 1036 | A B^T | 6822 -> 6893 (+71) GFLOPS
problemSize = 1037 | A B^T | 6785 -> 6848 (+63) GFLOPS
problemSize = 1038 | A B^T | 6877 -> 6938 (+61) GFLOPS
problemSize = 1039 | A B^T | 6891 -> 6955 (+64) GFLOPS
problemSize = 1040 | A B^T | 7047 -> 7086 (+39) GFLOPS
problemSize = 1041 | A B^T | 6475 -> 6517 (+42) GFLOPS
problemSize = 1042 | A B^T | 6612 -> 6636 (+24) GFLOPS
problemSize = 1043 | A B^T | 6636 -> 6680 (+44) GFLOPS
problemSize = 1044 | A B^T | 6805 -> 6802 (-3) GFLOPS
problemSize = 1045 | A B^T | 6820 -> 6851 (+31) GFLOPS
problemSize = 1046 | A B^T | 6908 -> 6912 (+4) GFLOPS
problemSize = 1047 | A B^T | 6948 -> 6931 (-17) GFLOPS
problemSize = 1048 | A B^T | 7073 -> 7029 (-44) GFLOPS
problemSize = 1049 | A B^T | 6804 -> 6884 (+80) GFLOPS
problemSize = 1050 | A B^T | 6901 -> 6991 (+90) GFLOPS
problemSize = 1051 | A B^T | 6948 -> 7023 (+75) GFLOPS
problemSize = 1052 | A B^T | 7129 -> 7163 (+34) GFLOPS
problemSize = 1053 | A B^T | 7107 -> 7186 (+79) GFLOPS
problemSize = 1054 | A B^T | 7223 -> 7296 (+73) GFLOPS
problemSize = 1055 | A B^T | 7245 -> 7295 (+50) GFLOPS
problemSize = 1056 | A B^T | 7371 -> 7394 (+23) GFLOPS
problemSize = 1057 | A B^T | 6534 -> 6633 (+99) GFLOPS
problemSize = 1058 | A B^T | 6680 -> 6747 (+67) GFLOPS
problemSize = 1059 | A B^T | 6715 -> 6785 (+70) GFLOPS
problemSize = 1060 | A B^T | 6891 -> 6951 (+60) GFLOPS
problemSize = 1061 | A B^T | 6860 -> 6912 (+52) GFLOPS
problemSize = 1062 | A B^T | 6961 -> 7001 (+40) GFLOPS
problemSize = 1063 | A B^T | 6970 -> 7027 (+57) GFLOPS
problemSize = 1064 | A B^T | 7133 -> 7150 (+17) GFLOPS
problemSize = 1065 | A B^T | 6603 -> 6626 (+23) GFLOPS
problemSize = 1066 | A B^T | 6696 -> 6726 (+30) GFLOPS
problemSize = 1067 | A B^T | 6742 -> 6769 (+27) GFLOPS
problemSize = 1068 | A B^T | 6880 -> 6891 (+11) GFLOPS
problemSize = 1069 | A B^T | 6881 -> 6895 (+14) GFLOPS
problemSize = 1070 | A B^T | 6977 -> 6957 (-20) GFLOPS
problemSize = 1071 | A B^T | 7014 -> 6998 (-16) GFLOPS
problemSize = 1072 | A B^T | 7128 -> 7089 (-39) GFLOPS
M1 Max, FP16
problemSize = 976 | A B | 8105 -> 8113 (+8) GFLOPS
problemSize = 977 | A B | 7950 -> 7947 (-3) GFLOPS
problemSize = 978 | A B | 7991 -> 7988 (-3) GFLOPS
problemSize = 979 | A B | 8023 -> 8023 ( 0) GFLOPS
problemSize = 980 | A B | 8065 -> 8071 (+6) GFLOPS
problemSize = 981 | A B | 8088 -> 8072 (-16) GFLOPS
problemSize = 982 | A B | 8127 -> 8118 (-9) GFLOPS
problemSize = 983 | A B | 8173 -> 8147 (-26) GFLOPS
problemSize = 984 | A B | 8181 -> 8179 (-2) GFLOPS
problemSize = 985 | A B | 8043 -> 8040 (-3) GFLOPS
problemSize = 986 | A B | 8089 -> 8106 (+17) GFLOPS
problemSize = 987 | A B | 8131 -> 8139 (+8) GFLOPS
problemSize = 988 | A B | 8166 -> 8148 (-18) GFLOPS
problemSize = 989 | A B | 8192 -> 8164 (-28) GFLOPS
problemSize = 990 | A B | 8216 -> 8198 (-18) GFLOPS
problemSize = 991 | A B | 8242 -> 8216 (-26) GFLOPS
problemSize = 992 | A B | 8368 -> 8324 (-44) GFLOPS
problemSize = 993 | A B | 8144 -> 8163 (+19) GFLOPS
problemSize = 994 | A B | 8178 -> 8221 (+43) GFLOPS
problemSize = 995 | A B | 8217 -> 8217 ( 0) GFLOPS
problemSize = 996 | A B | 8253 -> 8282 (+29) GFLOPS
problemSize = 997 | A B | 8275 -> 8269 (-6) GFLOPS
problemSize = 998 | A B | 8321 -> 8298 (-23) GFLOPS
problemSize = 999 | A B | 8332 -> 8329 (-3) GFLOPS
problemSize = 1000 | A B | 8393 -> 8382 (-11) GFLOPS
problemSize = 1001 | A B | 8349 -> 8346 (-3) GFLOPS
problemSize = 1002 | A B | 8388 -> 8392 (+4) GFLOPS
problemSize = 1003 | A B | 8419 -> 8420 (+1) GFLOPS
problemSize = 1004 | A B | 8458 -> 8460 (+2) GFLOPS
problemSize = 1005 | A B | 8473 -> 8504 (+31) GFLOPS
problemSize = 1006 | A B | 8509 -> 8508 (-1) GFLOPS
problemSize = 1007 | A B | 8534 -> 8564 (+30) GFLOPS
problemSize = 1008 | A B | 8572 -> 8598 (+26) GFLOPS
problemSize = 1009 | A B | 7872 -> 7857 (-15) GFLOPS
problemSize = 1010 | A B | 7905 -> 7914 (+9) GFLOPS
problemSize = 1011 | A B | 7933 -> 7920 (-13) GFLOPS
problemSize = 1012 | A B | 7971 -> 7947 (-24) GFLOPS
problemSize = 1013 | A B | 7999 -> 7990 (-9) GFLOPS
problemSize = 1014 | A B | 8026 -> 8007 (-19) GFLOPS
problemSize = 1015 | A B | 8068 -> 8030 (-38) GFLOPS
problemSize = 1016 | A B | 8088 -> 8078 (-10) GFLOPS
problemSize = 1017 | A B | 7895 -> 7747 (-148) GFLOPS
problemSize = 1018 | A B | 7931 -> 7766 (-165) GFLOPS
problemSize = 1019 | A B | 7938 -> 7798 (-140) GFLOPS
problemSize = 1020 | A B | 8005 -> 7833 (-172) GFLOPS
problemSize = 1021 | A B | 8004 -> 7894 (-110) GFLOPS
problemSize = 1022 | A B | 8044 -> 7883 (-161) GFLOPS
problemSize = 1023 | A B | 8053 -> 7925 (-128) GFLOPS
problemSize = 1024 | A B | 8124 -> 8103 (-21) GFLOPS
problemSize = 1025 | A B | 8090 -> 8095 (+5) GFLOPS
problemSize = 1026 | A B | 8117 -> 8123 (+6) GFLOPS
problemSize = 1027 | A B | 8177 -> 8143 (-34) GFLOPS
problemSize = 1028 | A B | 8184 -> 8178 (-6) GFLOPS
problemSize = 1029 | A B | 8227 -> 8204 (-23) GFLOPS
problemSize = 1030 | A B | 8225 -> 8225 ( 0) GFLOPS
problemSize = 1031 | A B | 8252 -> 8266 (+14) GFLOPS
problemSize = 1032 | A B | 8289 -> 8286 (-3) GFLOPS
problemSize = 1033 | A B | 7981 -> 8002 (+21) GFLOPS
problemSize = 1034 | A B | 8025 -> 8003 (-22) GFLOPS
problemSize = 1035 | A B | 8034 -> 8029 (-5) GFLOPS
problemSize = 1036 | A B | 8070 -> 8073 (+3) GFLOPS
problemSize = 1037 | A B | 8096 -> 8090 (-6) GFLOPS
problemSize = 1038 | A B | 8123 -> 8108 (-15) GFLOPS
problemSize = 1039 | A B | 8146 -> 8137 (-9) GFLOPS
problemSize = 1040 | A B | 8240 -> 8195 (-45) GFLOPS
problemSize = 1041 | A B | 8078 -> 8066 (-12) GFLOPS
problemSize = 1042 | A B | 8116 -> 8100 (-16) GFLOPS
problemSize = 1043 | A B | 8143 -> 8122 (-21) GFLOPS
problemSize = 1044 | A B | 8200 -> 8165 (-35) GFLOPS
problemSize = 1045 | A B | 8204 -> 8183 (-21) GFLOPS
problemSize = 1046 | A B | 8236 -> 8211 (-25) GFLOPS
problemSize = 1047 | A B | 8251 -> 8239 (-12) GFLOPS
problemSize = 1048 | A B | 8311 -> 8279 (-32) GFLOPS
problemSize = 1049 | A B | 8244 -> 8203 (-41) GFLOPS
problemSize = 1050 | A B | 8291 -> 8248 (-43) GFLOPS
problemSize = 1051 | A B | 8310 -> 8260 (-50) GFLOPS
problemSize = 1052 | A B | 8350 -> 8302 (-48) GFLOPS
problemSize = 1053 | A B | 8380 -> 8324 (-56) GFLOPS
problemSize = 1054 | A B | 8403 -> 8367 (-36) GFLOPS
problemSize = 1055 | A B | 8425 -> 8379 (-46) GFLOPS
problemSize = 1056 | A B | 8561 -> 8557 (-4) GFLOPS
problemSize = 1057 | A B | 7965 -> 7934 (-31) GFLOPS
problemSize = 1058 | A B | 8012 -> 7977 (-35) GFLOPS
problemSize = 1059 | A B | 8015 -> 7996 (-19) GFLOPS
problemSize = 1060 | A B | 8056 -> 8036 (-20) GFLOPS
problemSize = 1061 | A B | 8089 -> 8039 (-50) GFLOPS
problemSize = 1062 | A B | 8099 -> 8066 (-33) GFLOPS
problemSize = 1063 | A B | 8134 -> 8106 (-28) GFLOPS
problemSize = 1064 | A B | 8169 -> 8139 (-30) GFLOPS
problemSize = 1065 | A B | 8166 -> 8191 (+25) GFLOPS
problemSize = 1066 | A B | 8196 -> 8211 (+15) GFLOPS
problemSize = 1067 | A B | 8219 -> 8198 (-21) GFLOPS
problemSize = 1068 | A B | 8272 -> 8250 (-22) GFLOPS
problemSize = 1069 | A B | 8279 -> 8264 (-15) GFLOPS
problemSize = 1070 | A B | 8296 -> 8309 (+13) GFLOPS
problemSize = 1071 | A B | 8339 -> 8307 (-32) GFLOPS
problemSize = 1072 | A B | 8400 -> 8375 (-25) GFLOPS
problemSize = 976 | A B^T | 7962 -> 8012 (+50) GFLOPS
problemSize = 977 | A B^T | 7730 -> 7780 (+50) GFLOPS
problemSize = 978 | A B^T | 7830 -> 7878 (+48) GFLOPS
problemSize = 979 | A B^T | 7896 -> 7909 (+13) GFLOPS
problemSize = 980 | A B^T | 7962 -> 7971 (+9) GFLOPS
problemSize = 981 | A B^T | 8006 -> 7991 (-15) GFLOPS
problemSize = 982 | A B^T | 8052 -> 8028 (-24) GFLOPS
problemSize = 983 | A B^T | 8074 -> 8072 (-2) GFLOPS
problemSize = 984 | A B^T | 8136 -> 8086 (-50) GFLOPS
problemSize = 985 | A B^T | 7911 -> 7976 (+65) GFLOPS
problemSize = 986 | A B^T | 8063 -> 8088 (+25) GFLOPS
problemSize = 987 | A B^T | 8093 -> 8178 (+85) GFLOPS
problemSize = 988 | A B^T | 8192 -> 8233 (+41) GFLOPS
problemSize = 989 | A B^T | 8218 -> 8278 (+60) GFLOPS
problemSize = 990 | A B^T | 8292 -> 8314 (+22) GFLOPS
problemSize = 991 | A B^T | 8313 -> 8334 (+21) GFLOPS
problemSize = 992 | A B^T | 8485 -> 8497 (+12) GFLOPS
problemSize = 993 | A B^T | 7957 -> 8025 (+68) GFLOPS
problemSize = 994 | A B^T | 8044 -> 8106 (+62) GFLOPS
problemSize = 995 | A B^T | 8109 -> 8167 (+58) GFLOPS
problemSize = 996 | A B^T | 8199 -> 8228 (+29) GFLOPS
problemSize = 997 | A B^T | 8257 -> 8259 (+2) GFLOPS
problemSize = 998 | A B^T | 8273 -> 8299 (+26) GFLOPS
problemSize = 999 | A B^T | 8304 -> 8322 (+18) GFLOPS
problemSize = 1000 | A B^T | 8377 -> 8393 (+16) GFLOPS
problemSize = 1001 | A B^T | 8022 -> 8109 (+87) GFLOPS
problemSize = 1002 | A B^T | 8116 -> 8192 (+76) GFLOPS
problemSize = 1003 | A B^T | 8174 -> 8247 (+73) GFLOPS
problemSize = 1004 | A B^T | 8253 -> 8306 (+53) GFLOPS
problemSize = 1005 | A B^T | 8268 -> 8329 (+61) GFLOPS
problemSize = 1006 | A B^T | 8350 -> 8377 (+27) GFLOPS
problemSize = 1007 | A B^T | 8395 -> 8405 (+10) GFLOPS
problemSize = 1008 | A B^T | 8490 -> 8496 (+6) GFLOPS
problemSize = 1009 | A B^T | 7735 -> 7728 (-7) GFLOPS
problemSize = 1010 | A B^T | 7770 -> 7806 (+36) GFLOPS
problemSize = 1011 | A B^T | 7844 -> 7835 (-9) GFLOPS
problemSize = 1012 | A B^T | 7895 -> 7876 (-19) GFLOPS
problemSize = 1013 | A B^T | 7928 -> 7915 (-13) GFLOPS
problemSize = 1014 | A B^T | 7969 -> 7941 (-28) GFLOPS
problemSize = 1015 | A B^T | 7992 -> 7972 (-20) GFLOPS
problemSize = 1016 | A B^T | 8061 -> 8034 (-27) GFLOPS
problemSize = 1017 | A B^T | 7995 -> 8055 (+60) GFLOPS
problemSize = 1018 | A B^T | 8077 -> 8119 (+42) GFLOPS
problemSize = 1019 | A B^T | 8136 -> 8172 (+36) GFLOPS
problemSize = 1020 | A B^T | 8194 -> 8232 (+38) GFLOPS
problemSize = 1021 | A B^T | 8240 -> 8245 (+5) GFLOPS
problemSize = 1022 | A B^T | 8274 -> 8291 (+17) GFLOPS
problemSize = 1023 | A B^T | 8320 -> 8337 (+17) GFLOPS
problemSize = 1024 | A B^T | 8404 -> 8424 (+20) GFLOPS
problemSize = 1025 | A B^T | 8029 -> 8050 (+21) GFLOPS
problemSize = 1026 | A B^T | 8031 -> 8095 (+64) GFLOPS
problemSize = 1027 | A B^T | 8102 -> 8163 (+61) GFLOPS
problemSize = 1028 | A B^T | 8177 -> 8195 (+18) GFLOPS
problemSize = 1029 | A B^T | 8199 -> 8252 (+53) GFLOPS
problemSize = 1030 | A B^T | 8240 -> 8250 (+10) GFLOPS
problemSize = 1031 | A B^T | 8269 -> 8269 ( 0) GFLOPS
problemSize = 1032 | A B^T | 8343 -> 8345 (+2) GFLOPS
problemSize = 1033 | A B^T | 7736 -> 7815 (+79) GFLOPS
problemSize = 1034 | A B^T | 7808 -> 7883 (+75) GFLOPS
problemSize = 1035 | A B^T | 7853 -> 7910 (+57) GFLOPS
problemSize = 1036 | A B^T | 7904 -> 7964 (+60) GFLOPS
problemSize = 1037 | A B^T | 7935 -> 7993 (+58) GFLOPS
problemSize = 1038 | A B^T | 7976 -> 8035 (+59) GFLOPS
problemSize = 1039 | A B^T | 8007 -> 8075 (+68) GFLOPS
problemSize = 1040 | A B^T | 8069 -> 8133 (+64) GFLOPS
problemSize = 1041 | A B^T | 7908 -> 7929 (+21) GFLOPS
problemSize = 1042 | A B^T | 7985 -> 8008 (+23) GFLOPS
problemSize = 1043 | A B^T | 8037 -> 8030 (-7) GFLOPS
problemSize = 1044 | A B^T | 8114 -> 8092 (-22) GFLOPS
problemSize = 1045 | A B^T | 8133 -> 8111 (-22) GFLOPS
problemSize = 1046 | A B^T | 8173 -> 8151 (-22) GFLOPS
problemSize = 1047 | A B^T | 8200 -> 8171 (-29) GFLOPS
problemSize = 1048 | A B^T | 8264 -> 8260 (-4) GFLOPS
problemSize = 1049 | A B^T | 8191 -> 8261 (+70) GFLOPS
problemSize = 1050 | A B^T | 8266 -> 8319 (+53) GFLOPS
problemSize = 1051 | A B^T | 8318 -> 8372 (+54) GFLOPS
problemSize = 1052 | A B^T | 8407 -> 8435 (+28) GFLOPS
problemSize = 1053 | A B^T | 8436 -> 8471 (+35) GFLOPS
problemSize = 1054 | A B^T | 8479 -> 8497 (+18) GFLOPS
problemSize = 1055 | A B^T | 8517 -> 8530 (+13) GFLOPS
problemSize = 1056 | A B^T | 8623 -> 8625 (+2) GFLOPS
problemSize = 1057 | A B^T | 7838 -> 7901 (+63) GFLOPS
problemSize = 1058 | A B^T | 7939 -> 7971 (+32) GFLOPS
problemSize = 1059 | A B^T | 7958 -> 7999 (+41) GFLOPS
problemSize = 1060 | A B^T | 8022 -> 8040 (+18) GFLOPS
problemSize = 1061 | A B^T | 8066 -> 8039 (-27) GFLOPS
problemSize = 1062 | A B^T | 8091 -> 8123 (+32) GFLOPS
problemSize = 1063 | A B^T | 8136 -> 8142 (+6) GFLOPS
problemSize = 1064 | A B^T | 8190 -> 8146 (-44) GFLOPS
problemSize = 1065 | A B^T | 7943 -> 7970 (+27) GFLOPS
problemSize = 1066 | A B^T | 7969 -> 8022 (+53) GFLOPS
problemSize = 1067 | A B^T | 8026 -> 8093 (+67) GFLOPS
problemSize = 1068 | A B^T | 8116 -> 8131 (+15) GFLOPS
problemSize = 1069 | A B^T | 8104 -> 8152 (+48) GFLOPS
problemSize = 1070 | A B^T | 8143 -> 8182 (+39) GFLOPS
problemSize = 1071 | A B^T | 8214 -> 8197 (-17) GFLOPS
problemSize = 1072 | A B^T | 8244 -> 8282 (+38) GFLOPS
M1 Max, BF16
problemSize = 976 | A B | 7511 -> 7047 (-464) GFLOPS
problemSize = 977 | A B | 6904 -> 6889 (-15) GFLOPS
problemSize = 978 | A B | 6934 -> 6911 (-23) GFLOPS
problemSize = 979 | A B | 6974 -> 6957 (-17) GFLOPS
problemSize = 980 | A B | 6981 -> 6994 (+13) GFLOPS
problemSize = 981 | A B | 7020 -> 6998 (-22) GFLOPS
problemSize = 982 | A B | 7028 -> 7026 (-2) GFLOPS
problemSize = 983 | A B | 7061 -> 7052 (-9) GFLOPS
problemSize = 984 | A B | 7091 -> 7074 (-17) GFLOPS
problemSize = 985 | A B | 7000 -> 6885 (-115) GFLOPS
problemSize = 986 | A B | 7108 -> 6939 (-169) GFLOPS
problemSize = 987 | A B | 7093 -> 6942 (-151) GFLOPS
problemSize = 988 | A B | 7132 -> 6994 (-138) GFLOPS
problemSize = 989 | A B | 7124 -> 7018 (-106) GFLOPS
problemSize = 990 | A B | 7249 -> 7027 (-222) GFLOPS
problemSize = 991 | A B | 7183 -> 7044 (-139) GFLOPS
problemSize = 992 | A B | 7318 -> 7212 (-106) GFLOPS
problemSize = 993 | A B | 7167 -> 7155 (-12) GFLOPS
problemSize = 994 | A B | 7172 -> 7219 (+47) GFLOPS
problemSize = 995 | A B | 7183 -> 7220 (+37) GFLOPS
problemSize = 996 | A B | 7250 -> 7252 (+2) GFLOPS
problemSize = 997 | A B | 7270 -> 7268 (-2) GFLOPS
problemSize = 998 | A B | 7311 -> 7336 (+25) GFLOPS
problemSize = 999 | A B | 7305 -> 7294 (-11) GFLOPS
problemSize = 1000 | A B | 7366 -> 7387 (+21) GFLOPS
problemSize = 1001 | A B | 7787 -> 7119 (-668) GFLOPS
problemSize = 1002 | A B | 7826 -> 7181 (-645) GFLOPS
problemSize = 1003 | A B | 7858 -> 7213 (-645) GFLOPS
problemSize = 1004 | A B | 7896 -> 7222 (-674) GFLOPS
problemSize = 1005 | A B | 7922 -> 7230 (-692) GFLOPS
problemSize = 1006 | A B | 7939 -> 7295 (-644) GFLOPS
problemSize = 1007 | A B | 7966 -> 7310 (-656) GFLOPS
problemSize = 1008 | A B | 7989 -> 8022 (+33) GFLOPS
problemSize = 1009 | A B | 6685 -> 6673 (-12) GFLOPS
problemSize = 1010 | A B | 6716 -> 6708 (-8) GFLOPS
problemSize = 1011 | A B | 6739 -> 6718 (-21) GFLOPS
problemSize = 1012 | A B | 6808 -> 6771 (-37) GFLOPS
problemSize = 1013 | A B | 6796 -> 6779 (-17) GFLOPS
problemSize = 1014 | A B | 6818 -> 6804 (-14) GFLOPS
problemSize = 1015 | A B | 6832 -> 6836 (+4) GFLOPS
problemSize = 1016 | A B | 6882 -> 6870 (-12) GFLOPS
problemSize = 1017 | A B | 6980 -> 6709 (-271) GFLOPS
problemSize = 1018 | A B | 7038 -> 6722 (-316) GFLOPS
problemSize = 1019 | A B | 7035 -> 6746 (-289) GFLOPS
problemSize = 1020 | A B | 7075 -> 6773 (-302) GFLOPS
problemSize = 1021 | A B | 7101 -> 6788 (-313) GFLOPS
problemSize = 1022 | A B | 7130 -> 6832 (-298) GFLOPS
problemSize = 1023 | A B | 7129 -> 6845 (-284) GFLOPS
problemSize = 1024 | A B | 7250 -> 6931 (-319) GFLOPS
problemSize = 1025 | A B | 7051 -> 7078 (+27) GFLOPS
problemSize = 1026 | A B | 7114 -> 7085 (-29) GFLOPS
problemSize = 1027 | A B | 7111 -> 7127 (+16) GFLOPS
problemSize = 1028 | A B | 7168 -> 7144 (-24) GFLOPS
problemSize = 1029 | A B | 7136 -> 7150 (+14) GFLOPS
problemSize = 1030 | A B | 7218 -> 7186 (-32) GFLOPS
problemSize = 1031 | A B | 7196 -> 7200 (+4) GFLOPS
problemSize = 1032 | A B | 7251 -> 7254 (+3) GFLOPS
problemSize = 1033 | A B | 7031 -> 7051 (+20) GFLOPS
problemSize = 1034 | A B | 7064 -> 7083 (+19) GFLOPS
problemSize = 1035 | A B | 7070 -> 7108 (+38) GFLOPS
problemSize = 1036 | A B | 7138 -> 7146 (+8) GFLOPS
problemSize = 1037 | A B | 7121 -> 7157 (+36) GFLOPS
problemSize = 1038 | A B | 7163 -> 7186 (+23) GFLOPS
problemSize = 1039 | A B | 7177 -> 7208 (+31) GFLOPS
problemSize = 1040 | A B | 7259 -> 7275 (+16) GFLOPS
problemSize = 1041 | A B | 7106 -> 7087 (-19) GFLOPS
problemSize = 1042 | A B | 7106 -> 7114 (+8) GFLOPS
problemSize = 1043 | A B | 7143 -> 7134 (-9) GFLOPS
problemSize = 1044 | A B | 7186 -> 7166 (-20) GFLOPS
problemSize = 1045 | A B | 7192 -> 7185 (-7) GFLOPS
problemSize = 1046 | A B | 7210 -> 7214 (+4) GFLOPS
problemSize = 1047 | A B | 7268 -> 7244 (-24) GFLOPS
problemSize = 1048 | A B | 7273 -> 7288 (+15) GFLOPS
problemSize = 1049 | A B | 7412 -> 7028 (-384) GFLOPS
problemSize = 1050 | A B | 7456 -> 7065 (-391) GFLOPS
problemSize = 1051 | A B | 7450 -> 7094 (-356) GFLOPS
problemSize = 1052 | A B | 7546 -> 7123 (-423) GFLOPS
problemSize = 1053 | A B | 7517 -> 7142 (-375) GFLOPS
problemSize = 1054 | A B | 7573 -> 7165 (-408) GFLOPS
problemSize = 1055 | A B | 7582 -> 7179 (-403) GFLOPS
problemSize = 1056 | A B | 7696 -> 7692 (-4) GFLOPS
problemSize = 1057 | A B | 7106 -> 7092 (-14) GFLOPS
problemSize = 1058 | A B | 7139 -> 7138 (-1) GFLOPS
problemSize = 1059 | A B | 7129 -> 7142 (+13) GFLOPS
problemSize = 1060 | A B | 7180 -> 7178 (-2) GFLOPS
problemSize = 1061 | A B | 7197 -> 7184 (-13) GFLOPS
problemSize = 1062 | A B | 7233 -> 7219 (-14) GFLOPS
problemSize = 1063 | A B | 7224 -> 7246 (+22) GFLOPS
problemSize = 1064 | A B | 7317 -> 7303 (-14) GFLOPS
problemSize = 1065 | A B | 6875 -> 7168 (+293) GFLOPS
problemSize = 1066 | A B | 6952 -> 7195 (+243) GFLOPS
problemSize = 1067 | A B | 6893 -> 7215 (+322) GFLOPS
problemSize = 1068 | A B | 7002 -> 7280 (+278) GFLOPS
problemSize = 1069 | A B | 6925 -> 7263 (+338) GFLOPS
problemSize = 1070 | A B | 7027 -> 7298 (+271) GFLOPS
problemSize = 1071 | A B | 6981 -> 7301 (+320) GFLOPS
problemSize = 1072 | A B | 7088 -> 7388 (+300) GFLOPS
problemSize = 976 | A B^T | 7606 -> 7564 (-42) GFLOPS
problemSize = 977 | A B^T | 7418 -> 7061 (-357) GFLOPS
problemSize = 978 | A B^T | 7076 -> 7107 (+31) GFLOPS
problemSize = 979 | A B^T | 7515 -> 7158 (-357) GFLOPS
problemSize = 980 | A B^T | 7176 -> 7239 (+63) GFLOPS
problemSize = 981 | A B^T | 7596 -> 7243 (-353) GFLOPS
problemSize = 982 | A B^T | 7234 -> 7264 (+30) GFLOPS
problemSize = 983 | A B^T | 7681 -> 7288 (-393) GFLOPS
problemSize = 984 | A B^T | 7331 -> 7283 (-48) GFLOPS
problemSize = 985 | A B^T | 6860 -> 6799 (-61) GFLOPS
problemSize = 986 | A B^T | 6966 -> 6878 (-88) GFLOPS
problemSize = 987 | A B^T | 7001 -> 6888 (-113) GFLOPS
problemSize = 988 | A B^T | 7069 -> 6984 (-85) GFLOPS
problemSize = 989 | A B^T | 7054 -> 6967 (-87) GFLOPS
problemSize = 990 | A B^T | 7113 -> 7054 (-59) GFLOPS
problemSize = 991 | A B^T | 7160 -> 7026 (-134) GFLOPS
problemSize = 992 | A B^T | 7284 -> 7231 (-53) GFLOPS
problemSize = 993 | A B^T | 7070 -> 7297 (+227) GFLOPS
problemSize = 994 | A B^T | 7269 -> 7297 (+28) GFLOPS
problemSize = 995 | A B^T | 7194 -> 7389 (+195) GFLOPS
problemSize = 996 | A B^T | 7376 -> 7412 (+36) GFLOPS
problemSize = 997 | A B^T | 7249 -> 7434 (+185) GFLOPS
problemSize = 998 | A B^T | 7441 -> 7488 (+47) GFLOPS
problemSize = 999 | A B^T | 7313 -> 7518 (+205) GFLOPS
problemSize = 1000 | A B^T | 7575 -> 7580 (+5) GFLOPS
problemSize = 1001 | A B^T | 7669 -> 7692 (+23) GFLOPS
problemSize = 1002 | A B^T | 7773 -> 7757 (-16) GFLOPS
problemSize = 1003 | A B^T | 7787 -> 7798 (+11) GFLOPS
problemSize = 1004 | A B^T | 7899 -> 7852 (-47) GFLOPS
problemSize = 1005 | A B^T | 7885 -> 7881 (-4) GFLOPS
problemSize = 1006 | A B^T | 7956 -> 7915 (-41) GFLOPS
problemSize = 1007 | A B^T | 7939 -> 7937 (-2) GFLOPS
problemSize = 1008 | A B^T | 8085 -> 8096 (+11) GFLOPS
problemSize = 1009 | A B^T | 6821 -> 6885 (+64) GFLOPS
problemSize = 1010 | A B^T | 6899 -> 6936 (+37) GFLOPS
problemSize = 1011 | A B^T | 6918 -> 6959 (+41) GFLOPS
problemSize = 1012 | A B^T | 6971 -> 6993 (+22) GFLOPS
problemSize = 1013 | A B^T | 6980 -> 7026 (+46) GFLOPS
problemSize = 1014 | A B^T | 7030 -> 7082 (+52) GFLOPS
problemSize = 1015 | A B^T | 7032 -> 7091 (+59) GFLOPS
problemSize = 1016 | A B^T | 7116 -> 7138 (+22) GFLOPS
problemSize = 1017 | A B^T | 6923 -> 6615 (-308) GFLOPS
problemSize = 1018 | A B^T | 7030 -> 6653 (-377) GFLOPS
problemSize = 1019 | A B^T | 7002 -> 6709 (-293) GFLOPS
problemSize = 1020 | A B^T | 7104 -> 6743 (-361) GFLOPS
problemSize = 1021 | A B^T | 7093 -> 6772 (-321) GFLOPS
problemSize = 1022 | A B^T | 7177 -> 6798 (-379) GFLOPS
problemSize = 1023 | A B^T | 7127 -> 6821 (-306) GFLOPS
problemSize = 1024 | A B^T | 7358 -> 6958 (-400) GFLOPS
problemSize = 1025 | A B^T | 7029 -> 7221 (+192) GFLOPS
problemSize = 1026 | A B^T | 7155 -> 7202 (+47) GFLOPS
problemSize = 1027 | A B^T | 7108 -> 7259 (+151) GFLOPS
problemSize = 1028 | A B^T | 7262 -> 7295 (+33) GFLOPS
problemSize = 1029 | A B^T | 7168 -> 7293 (+125) GFLOPS
problemSize = 1030 | A B^T | 7310 -> 7322 (+12) GFLOPS
problemSize = 1031 | A B^T | 7257 -> 7328 (+71) GFLOPS
problemSize = 1032 | A B^T | 7326 -> 7379 (+53) GFLOPS
problemSize = 1033 | A B^T | 7381 -> 7040 (-341) GFLOPS
problemSize = 1034 | A B^T | 7454 -> 7084 (-370) GFLOPS
problemSize = 1035 | A B^T | 7476 -> 7118 (-358) GFLOPS
problemSize = 1036 | A B^T | 7546 -> 7162 (-384) GFLOPS
problemSize = 1037 | A B^T | 7552 -> 7178 (-374) GFLOPS
problemSize = 1038 | A B^T | 7616 -> 7221 (-395) GFLOPS
problemSize = 1039 | A B^T | 7630 -> 7232 (-398) GFLOPS
problemSize = 1040 | A B^T | 7693 -> 7330 (-363) GFLOPS
problemSize = 1041 | A B^T | 7194 -> 7265 (+71) GFLOPS
problemSize = 1042 | A B^T | 7281 -> 7318 (+37) GFLOPS
problemSize = 1043 | A B^T | 7288 -> 7363 (+75) GFLOPS
problemSize = 1044 | A B^T | 7375 -> 7408 (+33) GFLOPS
problemSize = 1045 | A B^T | 7359 -> 7440 (+81) GFLOPS
problemSize = 1046 | A B^T | 7431 -> 7476 (+45) GFLOPS
problemSize = 1047 | A B^T | 7415 -> 7485 (+70) GFLOPS
problemSize = 1048 | A B^T | 7517 -> 7552 (+35) GFLOPS
problemSize = 1049 | A B^T | 7319 -> 6940 (-379) GFLOPS
problemSize = 1050 | A B^T | 7414 -> 6986 (-428) GFLOPS
problemSize = 1051 | A B^T | 7414 -> 7034 (-380) GFLOPS
problemSize = 1052 | A B^T | 7507 -> 7107 (-400) GFLOPS
problemSize = 1053 | A B^T | 7517 -> 7110 (-407) GFLOPS
problemSize = 1054 | A B^T | 7563 -> 7148 (-415) GFLOPS
problemSize = 1055 | A B^T | 7555 -> 7188 (-367) GFLOPS
problemSize = 1056 | A B^T | 7721 -> 7731 (+10) GFLOPS
problemSize = 1057 | A B^T | 7073 -> 7216 (+143) GFLOPS
problemSize = 1058 | A B^T | 7217 -> 7246 (+29) GFLOPS
problemSize = 1059 | A B^T | 7152 -> 7288 (+136) GFLOPS
problemSize = 1060 | A B^T | 7300 -> 7308 (+8) GFLOPS
problemSize = 1061 | A B^T | 7228 -> 7335 (+107) GFLOPS
problemSize = 1062 | A B^T | 7349 -> 7386 (+37) GFLOPS
problemSize = 1063 | A B^T | 7280 -> 7401 (+121) GFLOPS
problemSize = 1064 | A B^T | 7428 -> 7433 (+5) GFLOPS
problemSize = 1065 | A B^T | 6943 -> 7013 (+70) GFLOPS
problemSize = 1066 | A B^T | 7033 -> 7058 (+25) GFLOPS
problemSize = 1067 | A B^T | 7011 -> 7099 (+88) GFLOPS
problemSize = 1068 | A B^T | 7150 -> 7135 (-15) GFLOPS
problemSize = 1069 | A B^T | 7072 -> 7139 (+67) GFLOPS
problemSize = 1070 | A B^T | 7145 -> 7177 (+32) GFLOPS
problemSize = 1071 | A B^T | 7124 -> 7184 (+60) GFLOPS
problemSize = 1072 | A B^T | 7250 -> 7271 (+21) GFLOPS
M4, FP32
problemSize = 976 | A B | 2957 -> 2944 (-13) GFLOPS
problemSize = 977 | A B | 2943 -> 2934 (-9) GFLOPS
problemSize = 978 | A B | 2943 -> 2952 (+9) GFLOPS
problemSize = 979 | A B | 2948 -> 2953 (+5) GFLOPS
problemSize = 980 | A B | 2967 -> 2980 (+13) GFLOPS
problemSize = 981 | A B | 2967 -> 2970 (+3) GFLOPS
problemSize = 982 | A B | 2979 -> 2984 (+5) GFLOPS
problemSize = 983 | A B | 2983 -> 2991 (+8) GFLOPS
problemSize = 984 | A B | 3001 -> 2996 (-5) GFLOPS
problemSize = 985 | A B | 2986 -> 2985 (-1) GFLOPS
problemSize = 986 | A B | 2997 -> 3000 (+3) GFLOPS
problemSize = 987 | A B | 3004 -> 2989 (-15) GFLOPS
problemSize = 988 | A B | 3021 -> 3026 (+5) GFLOPS
problemSize = 989 | A B | 3010 -> 3014 (+4) GFLOPS
problemSize = 990 | A B | 3030 -> 3034 (+4) GFLOPS
problemSize = 991 | A B | 3029 -> 3032 (+3) GFLOPS
problemSize = 992 | A B | 3040 -> 3049 (+9) GFLOPS
problemSize = 993 | A B | 2860 -> 2864 (+4) GFLOPS
problemSize = 994 | A B | 2870 -> 2879 (+9) GFLOPS
problemSize = 995 | A B | 2875 -> 2879 (+4) GFLOPS
problemSize = 996 | A B | 2892 -> 2904 (+12) GFLOPS
problemSize = 997 | A B | 2886 -> 2900 (+14) GFLOPS
problemSize = 998 | A B | 2899 -> 2910 (+11) GFLOPS
problemSize = 999 | A B | 2904 -> 2907 (+3) GFLOPS
problemSize = 1000 | A B | 2914 -> 2920 (+6) GFLOPS
problemSize = 1001 | A B | 2902 -> 2900 (-2) GFLOPS
problemSize = 1002 | A B | 2914 -> 2922 (+8) GFLOPS
problemSize = 1003 | A B | 2917 -> 2923 (+6) GFLOPS
problemSize = 1004 | A B | 2939 -> 2947 (+8) GFLOPS
problemSize = 1005 | A B | 2929 -> 2930 (+1) GFLOPS
problemSize = 1006 | A B | 2937 -> 2947 (+10) GFLOPS
problemSize = 1007 | A B | 2954 -> 2947 (-7) GFLOPS
problemSize = 1008 | A B | 2952 -> 2968 (+16) GFLOPS
problemSize = 1009 | A B | 2947 -> 2942 (-5) GFLOPS
problemSize = 1010 | A B | 2952 -> 2961 (+9) GFLOPS
problemSize = 1011 | A B | 2950 -> 2957 (+7) GFLOPS
problemSize = 1012 | A B | 2988 -> 2992 (+4) GFLOPS
problemSize = 1013 | A B | 2965 -> 2962 (-3) GFLOPS
problemSize = 1014 | A B | 2977 -> 2982 (+5) GFLOPS
problemSize = 1015 | A B | 2962 -> 2963 (+1) GFLOPS
problemSize = 1016 | A B | 3003 -> 3008 (+5) GFLOPS
problemSize = 1017 | A B | 2959 -> 2954 (-5) GFLOPS
problemSize = 1018 | A B | 2989 -> 2987 (-2) GFLOPS
problemSize = 1019 | A B | 2972 -> 2965 (-7) GFLOPS
problemSize = 1020 | A B | 3025 -> 3038 (+13) GFLOPS
problemSize = 1021 | A B | 2977 -> 2970 (-7) GFLOPS
problemSize = 1022 | A B | 3018 -> 3012 (-6) GFLOPS
problemSize = 1023 | A B | 3011 -> 3013 (+2) GFLOPS
problemSize = 1024 | A B | 3056 -> 3062 (+6) GFLOPS
problemSize = 1025 | A B | 2806 -> 2835 (+29) GFLOPS
problemSize = 1026 | A B | 2845 -> 2813 (-32) GFLOPS
problemSize = 1027 | A B | 2839 -> 2825 (-14) GFLOPS
problemSize = 1028 | A B | 2913 -> 2925 (+12) GFLOPS
problemSize = 1029 | A B | 2844 -> 2840 (-4) GFLOPS
problemSize = 1030 | A B | 2891 -> 2880 (-11) GFLOPS
problemSize = 1031 | A B | 2857 -> 2845 (-12) GFLOPS
problemSize = 1032 | A B | 2948 -> 2940 (-8) GFLOPS
problemSize = 1033 | A B | 2841 -> 2836 (-5) GFLOPS
problemSize = 1034 | A B | 2901 -> 2893 (-8) GFLOPS
problemSize = 1035 | A B | 2869 -> 2855 (-14) GFLOPS
problemSize = 1036 | A B | 2957 -> 2967 (+10) GFLOPS
problemSize = 1037 | A B | 2887 -> 2872 (-15) GFLOPS
problemSize = 1038 | A B | 2935 -> 2919 (-16) GFLOPS
problemSize = 1039 | A B | 2916 -> 2907 (-9) GFLOPS
problemSize = 1040 | A B | 3003 -> 2992 (-11) GFLOPS
problemSize = 1041 | A B | 2904 -> 2904 ( 0) GFLOPS
problemSize = 1042 | A B | 2955 -> 2940 (-15) GFLOPS
problemSize = 1043 | A B | 2914 -> 2912 (-2) GFLOPS
problemSize = 1044 | A B | 3005 -> 3007 (+2) GFLOPS
problemSize = 1045 | A B | 2924 -> 2918 (-6) GFLOPS
problemSize = 1046 | A B | 2971 -> 2963 (-8) GFLOPS
problemSize = 1047 | A B | 2939 -> 2915 (-24) GFLOPS
problemSize = 1048 | A B | 3047 -> 3025 (-22) GFLOPS
problemSize = 1049 | A B | 2934 -> 2909 (-25) GFLOPS
problemSize = 1050 | A B | 2981 -> 2970 (-11) GFLOPS
problemSize = 1051 | A B | 2943 -> 2934 (-9) GFLOPS
problemSize = 1052 | A B | 3045 -> 3057 (+12) GFLOPS
problemSize = 1053 | A B | 2960 -> 2947 (-13) GFLOPS
problemSize = 1054 | A B | 3011 -> 2999 (-12) GFLOPS
problemSize = 1055 | A B | 2977 -> 2967 (-10) GFLOPS
problemSize = 1056 | A B | 3097 -> 3082 (-15) GFLOPS
problemSize = 1057 | A B | 2817 -> 2812 (-5) GFLOPS
problemSize = 1058 | A B | 2877 -> 2864 (-13) GFLOPS
problemSize = 1059 | A B | 2822 -> 2818 (-4) GFLOPS
problemSize = 1060 | A B | 2936 -> 2930 (-6) GFLOPS
problemSize = 1061 | A B | 2839 -> 2837 (-2) GFLOPS
problemSize = 1062 | A B | 2902 -> 2899 (-3) GFLOPS
problemSize = 1063 | A B | 2843 -> 2847 (+4) GFLOPS
problemSize = 1064 | A B | 2943 -> 2968 (+25) GFLOPS
problemSize = 1065 | A B | 2816 -> 2825 (+9) GFLOPS
problemSize = 1066 | A B | 2903 -> 2898 (-5) GFLOPS
problemSize = 1067 | A B | 2870 -> 2844 (-26) GFLOPS
problemSize = 1068 | A B | 2988 -> 2976 (-12) GFLOPS
problemSize = 1069 | A B | 2875 -> 2855 (-20) GFLOPS
problemSize = 1070 | A B | 2904 -> 2913 (+9) GFLOPS
problemSize = 1071 | A B | 2891 -> 2889 (-2) GFLOPS
problemSize = 1072 | A B | 2997 -> 3001 (+4) GFLOPS
problemSize = 976 | A B^T | 2908 -> 2908 ( 0) GFLOPS
problemSize = 977 | A B^T | 2609 -> 2630 (+21) GFLOPS
problemSize = 978 | A B^T | 2768 -> 2774 (+6) GFLOPS
problemSize = 979 | A B^T | 2628 -> 2656 (+28) GFLOPS
problemSize = 980 | A B^T | 2917 -> 2924 (+7) GFLOPS
problemSize = 981 | A B^T | 2657 -> 2662 (+5) GFLOPS
problemSize = 982 | A B^T | 2812 -> 2814 (+2) GFLOPS
problemSize = 983 | A B^T | 2700 -> 2714 (+14) GFLOPS
problemSize = 984 | A B^T | 2952 -> 2953 (+1) GFLOPS
problemSize = 985 | A B^T | 2695 -> 2723 (+28) GFLOPS
problemSize = 986 | A B^T | 2860 -> 2862 (+2) GFLOPS
problemSize = 987 | A B^T | 2695 -> 2705 (+10) GFLOPS
problemSize = 988 | A B^T | 2964 -> 2968 (+4) GFLOPS
problemSize = 989 | A B^T | 2711 -> 2728 (+17) GFLOPS
problemSize = 990 | A B^T | 2856 -> 2860 (+4) GFLOPS
problemSize = 991 | A B^T | 2709 -> 2736 (+27) GFLOPS
problemSize = 992 | A B^T | 3013 -> 3006 (-7) GFLOPS
problemSize = 993 | A B^T | 2562 -> 2583 (+21) GFLOPS
problemSize = 994 | A B^T | 2712 -> 2731 (+19) GFLOPS
problemSize = 995 | A B^T | 2591 -> 2617 (+26) GFLOPS
problemSize = 996 | A B^T | 2838 -> 2845 (+7) GFLOPS
problemSize = 997 | A B^T | 2624 -> 2650 (+26) GFLOPS
problemSize = 998 | A B^T | 2768 -> 2782 (+14) GFLOPS
problemSize = 999 | A B^T | 2670 -> 2692 (+22) GFLOPS
problemSize = 1000 | A B^T | 2858 -> 2879 (+21) GFLOPS
problemSize = 1001 | A B^T | 2693 -> 2712 (+19) GFLOPS
problemSize = 1002 | A B^T | 2808 -> 2823 (+15) GFLOPS
problemSize = 1003 | A B^T | 2706 -> 2707 (+1) GFLOPS
problemSize = 1004 | A B^T | 2879 -> 2890 (+11) GFLOPS
problemSize = 1005 | A B^T | 2676 -> 2694 (+18) GFLOPS
problemSize = 1006 | A B^T | 2797 -> 2809 (+12) GFLOPS
problemSize = 1007 | A B^T | 2685 -> 2705 (+20) GFLOPS
problemSize = 1008 | A B^T | 2929 -> 2923 (-6) GFLOPS
problemSize = 1009 | A B^T | 2703 -> 2707 (+4) GFLOPS
problemSize = 1010 | A B^T | 2838 -> 2837 (-1) GFLOPS
problemSize = 1011 | A B^T | 2734 -> 2748 (+14) GFLOPS
problemSize = 1012 | A B^T | 2921 -> 2930 (+9) GFLOPS
problemSize = 1013 | A B^T | 2704 -> 2711 (+7) GFLOPS
problemSize = 1014 | A B^T | 2833 -> 2839 (+6) GFLOPS
problemSize = 1015 | A B^T | 2678 -> 2699 (+21) GFLOPS
problemSize = 1016 | A B^T | 2955 -> 2964 (+9) GFLOPS
problemSize = 1017 | A B^T | 2677 -> 2695 (+18) GFLOPS
problemSize = 1018 | A B^T | 2846 -> 2851 (+5) GFLOPS
problemSize = 1019 | A B^T | 2692 -> 2705 (+13) GFLOPS
problemSize = 1020 | A B^T | 2963 -> 2966 (+3) GFLOPS
problemSize = 1021 | A B^T | 2711 -> 2722 (+11) GFLOPS
problemSize = 1022 | A B^T | 2856 -> 2859 (+3) GFLOPS
problemSize = 1023 | A B^T | 2754 -> 2786 (+32) GFLOPS
problemSize = 1024 | A B^T | 3035 -> 3027 (-8) GFLOPS
problemSize = 1025 | A B^T | 2615 -> 2617 (+2) GFLOPS
problemSize = 1026 | A B^T | 2762 -> 2756 (-6) GFLOPS
problemSize = 1027 | A B^T | 2623 -> 2623 ( 0) GFLOPS
problemSize = 1028 | A B^T | 2846 -> 2855 (+9) GFLOPS
problemSize = 1029 | A B^T | 2582 -> 2582 ( 0) GFLOPS
problemSize = 1030 | A B^T | 2738 -> 2738 ( 0) GFLOPS
problemSize = 1031 | A B^T | 2615 -> 2634 (+19) GFLOPS
problemSize = 1032 | A B^T | 2899 -> 2894 (-5) GFLOPS
problemSize = 1033 | A B^T | 2627 -> 2640 (+13) GFLOPS
problemSize = 1034 | A B^T | 2785 -> 2791 (+6) GFLOPS
problemSize = 1035 | A B^T | 2680 -> 2688 (+8) GFLOPS
problemSize = 1036 | A B^T | 2894 -> 2903 (+9) GFLOPS
problemSize = 1037 | A B^T | 2712 -> 2720 (+8) GFLOPS
problemSize = 1038 | A B^T | 2812 -> 2812 ( 0) GFLOPS
problemSize = 1039 | A B^T | 2697 -> 2706 (+9) GFLOPS
problemSize = 1040 | A B^T | 2953 -> 2949 (-4) GFLOPS
problemSize = 1041 | A B^T | 2722 -> 2717 (-5) GFLOPS
problemSize = 1042 | A B^T | 2832 -> 2837 (+5) GFLOPS
problemSize = 1043 | A B^T | 2730 -> 2738 (+8) GFLOPS
problemSize = 1044 | A B^T | 2935 -> 2945 (+10) GFLOPS
problemSize = 1045 | A B^T | 2752 -> 2768 (+16) GFLOPS
problemSize = 1046 | A B^T | 2878 -> 2867 (-11) GFLOPS
problemSize = 1047 | A B^T | 2759 -> 2770 (+11) GFLOPS
problemSize = 1048 | A B^T | 2990 -> 2986 (-4) GFLOPS
problemSize = 1049 | A B^T | 2774 -> 2774 ( 0) GFLOPS
problemSize = 1050 | A B^T | 2869 -> 2870 (+1) GFLOPS
problemSize = 1051 | A B^T | 2752 -> 2746 (-6) GFLOPS
problemSize = 1052 | A B^T | 2974 -> 2977 (+3) GFLOPS
problemSize = 1053 | A B^T | 2722 -> 2731 (+9) GFLOPS
problemSize = 1054 | A B^T | 2891 -> 2874 (-17) GFLOPS
problemSize = 1055 | A B^T | 2726 -> 2729 (+3) GFLOPS
problemSize = 1056 | A B^T | 3047 -> 3031 (-16) GFLOPS
problemSize = 1057 | A B^T | 2595 -> 2584 (-11) GFLOPS
problemSize = 1058 | A B^T | 2719 -> 2720 (+1) GFLOPS
problemSize = 1059 | A B^T | 2589 -> 2614 (+25) GFLOPS
problemSize = 1060 | A B^T | 2866 -> 2857 (-9) GFLOPS
problemSize = 1061 | A B^T | 2604 -> 2618 (+14) GFLOPS
problemSize = 1062 | A B^T | 2757 -> 2765 (+8) GFLOPS
problemSize = 1063 | A B^T | 2641 -> 2650 (+9) GFLOPS
problemSize = 1064 | A B^T | 2898 -> 2919 (+21) GFLOPS
problemSize = 1065 | A B^T | 2649 -> 2654 (+5) GFLOPS
problemSize = 1066 | A B^T | 2778 -> 2783 (+5) GFLOPS
problemSize = 1067 | A B^T | 2338 -> 2581 (+243) GFLOPS
problemSize = 1068 | A B^T | 2421 -> 2485 (+64) GFLOPS
problemSize = 1069 | A B^T | 2072 -> 2156 (+84) GFLOPS
problemSize = 1070 | A B^T | 2174 -> 2205 (+31) GFLOPS
problemSize = 1071 | A B^T | 2082 -> 2068 (-14) GFLOPS
problemSize = 1072 | A B^T | 2327 -> 2365 (+38) GFLOPS
M4, FP16
problemSize = 976 | A B | 3289 -> 3293 (+4) GFLOPS
problemSize = 977 | A B | 3251 -> 3250 (-1) GFLOPS
problemSize = 978 | A B | 3268 -> 3263 (-5) GFLOPS
problemSize = 979 | A B | 3273 -> 3263 (-10) GFLOPS
problemSize = 980 | A B | 3306 -> 3295 (-11) GFLOPS
problemSize = 981 | A B | 3288 -> 3272 (-16) GFLOPS
problemSize = 982 | A B | 3299 -> 3286 (-13) GFLOPS
problemSize = 983 | A B | 3311 -> 3285 (-26) GFLOPS
problemSize = 984 | A B | 3342 -> 3345 (+3) GFLOPS
problemSize = 985 | A B | 3291 -> 3282 (-9) GFLOPS
problemSize = 986 | A B | 3321 -> 3302 (-19) GFLOPS
problemSize = 987 | A B | 3307 -> 3297 (-10) GFLOPS
problemSize = 988 | A B | 3355 -> 3341 (-14) GFLOPS
problemSize = 989 | A B | 3331 -> 3311 (-20) GFLOPS
problemSize = 990 | A B | 3356 -> 3333 (-23) GFLOPS
problemSize = 991 | A B | 3358 -> 3334 (-24) GFLOPS
problemSize = 992 | A B | 3401 -> 3386 (-15) GFLOPS
problemSize = 993 | A B | 3146 -> 3149 (+3) GFLOPS
problemSize = 994 | A B | 3169 -> 3167 (-2) GFLOPS
problemSize = 995 | A B | 3164 -> 3154 (-10) GFLOPS
problemSize = 996 | A B | 3207 -> 3197 (-10) GFLOPS
problemSize = 997 | A B | 3173 -> 3169 (-4) GFLOPS
problemSize = 998 | A B | 3200 -> 3198 (-2) GFLOPS
problemSize = 999 | A B | 3187 -> 3179 (-8) GFLOPS
problemSize = 1000 | A B | 3244 -> 3249 (+5) GFLOPS
problemSize = 1001 | A B | 3168 -> 3176 (+8) GFLOPS
problemSize = 1002 | A B | 3204 -> 3205 (+1) GFLOPS
problemSize = 1003 | A B | 3195 -> 3186 (-9) GFLOPS
problemSize = 1004 | A B | 3246 -> 3245 (-1) GFLOPS
problemSize = 1005 | A B | 3207 -> 3198 (-9) GFLOPS
problemSize = 1006 | A B | 3237 -> 3232 (-5) GFLOPS
problemSize = 1007 | A B | 3222 -> 3213 (-9) GFLOPS
problemSize = 1008 | A B | 3297 -> 3299 (+2) GFLOPS
problemSize = 1009 | A B | 3208 -> 3203 (-5) GFLOPS
problemSize = 1010 | A B | 3246 -> 3226 (-20) GFLOPS
problemSize = 1011 | A B | 3230 -> 3217 (-13) GFLOPS
problemSize = 1012 | A B | 3296 -> 3289 (-7) GFLOPS
problemSize = 1013 | A B | 3242 -> 3236 (-6) GFLOPS
problemSize = 1014 | A B | 3275 -> 3271 (-4) GFLOPS
problemSize = 1015 | A B | 3256 -> 3252 (-4) GFLOPS
problemSize = 1016 | A B | 3350 -> 3349 (-1) GFLOPS
problemSize = 1017 | A B | 3243 -> 3239 (-4) GFLOPS
problemSize = 1018 | A B | 3288 -> 3270 (-18) GFLOPS
problemSize = 1019 | A B | 3263 -> 3257 (-6) GFLOPS
problemSize = 1020 | A B | 3343 -> 3333 (-10) GFLOPS
problemSize = 1021 | A B | 3277 -> 3266 (-11) GFLOPS
problemSize = 1022 | A B | 3328 -> 3304 (-24) GFLOPS
problemSize = 1023 | A B | 3303 -> 3284 (-19) GFLOPS
problemSize = 1024 | A B | 3412 -> 3397 (-15) GFLOPS
problemSize = 1025 | A B | 3128 -> 3114 (-14) GFLOPS
problemSize = 1026 | A B | 3164 -> 3143 (-21) GFLOPS
problemSize = 1027 | A B | 3131 -> 3118 (-13) GFLOPS
problemSize = 1028 | A B | 3207 -> 3207 ( 0) GFLOPS
problemSize = 1029 | A B | 3146 -> 3133 (-13) GFLOPS
problemSize = 1030 | A B | 3182 -> 3170 (-12) GFLOPS
problemSize = 1031 | A B | 3165 -> 3149 (-16) GFLOPS
problemSize = 1032 | A B | 3292 -> 3278 (-14) GFLOPS
problemSize = 1033 | A B | 3152 -> 3141 (-11) GFLOPS
problemSize = 1034 | A B | 3186 -> 3178 (-8) GFLOPS
problemSize = 1035 | A B | 3169 -> 3149 (-20) GFLOPS
problemSize = 1036 | A B | 3253 -> 3251 (-2) GFLOPS
problemSize = 1037 | A B | 3177 -> 3174 (-3) GFLOPS
problemSize = 1038 | A B | 3217 -> 3205 (-12) GFLOPS
problemSize = 1039 | A B | 3197 -> 3186 (-11) GFLOPS
problemSize = 1040 | A B | 3346 -> 3332 (-14) GFLOPS
problemSize = 1041 | A B | 3194 -> 3178 (-16) GFLOPS
problemSize = 1042 | A B | 3231 -> 3213 (-18) GFLOPS
problemSize = 1043 | A B | 3210 -> 3194 (-16) GFLOPS
problemSize = 1044 | A B | 3296 -> 3282 (-14) GFLOPS
problemSize = 1045 | A B | 3229 -> 3216 (-13) GFLOPS
problemSize = 1046 | A B | 3270 -> 3254 (-16) GFLOPS
problemSize = 1047 | A B | 3243 -> 3221 (-22) GFLOPS
problemSize = 1048 | A B | 3391 -> 3374 (-17) GFLOPS
problemSize = 1049 | A B | 3233 -> 3217 (-16) GFLOPS
problemSize = 1050 | A B | 3278 -> 3252 (-26) GFLOPS
problemSize = 1051 | A B | 3180 -> 3242 (+62) GFLOPS
problemSize = 1052 | A B | 3009 -> 3057 (+48) GFLOPS
problemSize = 1053 | A B | 2883 -> 2933 (+50) GFLOPS
problemSize = 1054 | A B | 2753 -> 2783 (+30) GFLOPS
problemSize = 1055 | A B | 2743 -> 2719 (-24) GFLOPS
problemSize = 1056 | A B | 2777 -> 2717 (-60) GFLOPS
problemSize = 1057 | A B | 2670 -> 2658 (-12) GFLOPS
problemSize = 1058 | A B | 2694 -> 2650 (-44) GFLOPS
problemSize = 1059 | A B | 2702 -> 2674 (-28) GFLOPS
problemSize = 1060 | A B | 2698 -> 2688 (-10) GFLOPS
problemSize = 1061 | A B | 2727 -> 2706 (-21) GFLOPS
problemSize = 1062 | A B | 2722 -> 2706 (-16) GFLOPS
problemSize = 1063 | A B | 2724 -> 2699 (-25) GFLOPS
problemSize = 1064 | A B | 2716 -> 2733 (+17) GFLOPS
problemSize = 1065 | A B | 2758 -> 2767 (+9) GFLOPS
problemSize = 1066 | A B | 2759 -> 2756 (-3) GFLOPS
problemSize = 1067 | A B | 2749 -> 2756 (+7) GFLOPS
problemSize = 1068 | A B | 2764 -> 2757 (-7) GFLOPS
problemSize = 1069 | A B | 2780 -> 2744 (-36) GFLOPS
problemSize = 1070 | A B | 2781 -> 2751 (-30) GFLOPS
problemSize = 1071 | A B | 2815 -> 2815 ( 0) GFLOPS
problemSize = 1072 | A B | 2707 -> 2747 (+40) GFLOPS
problemSize = 977 | A B^T | 3184 -> 3184 ( 0) GFLOPS
problemSize = 978 | A B^T | 3218 -> 3205 (-13) GFLOPS
problemSize = 979 | A B^T | 3193 -> 3184 (-9) GFLOPS
problemSize = 980 | A B^T | 3253 -> 3236 (-17) GFLOPS
problemSize = 981 | A B^T | 3205 -> 3190 (-15) GFLOPS
problemSize = 982 | A B^T | 3216 -> 3213 (-3) GFLOPS
problemSize = 983 | A B^T | 3214 -> 3199 (-15) GFLOPS
problemSize = 984 | A B^T | 3328 -> 3306 (-22) GFLOPS
problemSize = 985 | A B^T | 3206 -> 3195 (-11) GFLOPS
problemSize = 986 | A B^T | 3223 -> 3218 (-5) GFLOPS
problemSize = 987 | A B^T | 3225 -> 3195 (-30) GFLOPS
problemSize = 988 | A B^T | 3276 -> 3269 (-7) GFLOPS
problemSize = 989 | A B^T | 3238 -> 3218 (-20) GFLOPS
problemSize = 990 | A B^T | 3262 -> 3238 (-24) GFLOPS
problemSize = 991 | A B^T | 3252 -> 3236 (-16) GFLOPS
problemSize = 992 | A B^T | 3391 -> 3372 (-19) GFLOPS
problemSize = 993 | A B^T | 3048 -> 3041 (-7) GFLOPS
problemSize = 994 | A B^T | 3066 -> 3071 (+5) GFLOPS
problemSize = 995 | A B^T | 3045 -> 3048 (+3) GFLOPS
problemSize = 996 | A B^T | 3108 -> 3114 (+6) GFLOPS
problemSize = 997 | A B^T | 3058 -> 3055 (-3) GFLOPS
problemSize = 998 | A B^T | 3083 -> 3087 (+4) GFLOPS
problemSize = 999 | A B^T | 3083 -> 3074 (-9) GFLOPS
problemSize = 1000 | A B^T | 3224 -> 3207 (-17) GFLOPS
problemSize = 1001 | A B^T | 3071 -> 3067 (-4) GFLOPS
problemSize = 1002 | A B^T | 3097 -> 3089 (-8) GFLOPS
problemSize = 1003 | A B^T | 3082 -> 3077 (-5) GFLOPS
problemSize = 1004 | A B^T | 3139 -> 3138 (-1) GFLOPS
problemSize = 1005 | A B^T | 3096 -> 3090 (-6) GFLOPS
problemSize = 1006 | A B^T | 3129 -> 3113 (-16) GFLOPS
problemSize = 1007 | A B^T | 3119 -> 3112 (-7) GFLOPS
problemSize = 1008 | A B^T | 3287 -> 3262 (-25) GFLOPS
problemSize = 1009 | A B^T | 3116 -> 3103 (-13) GFLOPS
problemSize = 1010 | A B^T | 3134 -> 3132 (-2) GFLOPS
problemSize = 1011 | A B^T | 3130 -> 3118 (-12) GFLOPS
problemSize = 1012 | A B^T | 3175 -> 3171 (-4) GFLOPS
problemSize = 1013 | A B^T | 3138 -> 3130 (-8) GFLOPS
problemSize = 1014 | A B^T | 3167 -> 3153 (-14) GFLOPS
problemSize = 1015 | A B^T | 3150 -> 3145 (-5) GFLOPS
problemSize = 1016 | A B^T | 3327 -> 3287 (-40) GFLOPS
problemSize = 1017 | A B^T | 3150 -> 3139 (-11) GFLOPS
problemSize = 1018 | A B^T | 3172 -> 3161 (-11) GFLOPS
problemSize = 1019 | A B^T | 3165 -> 3153 (-12) GFLOPS
problemSize = 1020 | A B^T | 3213 -> 3201 (-12) GFLOPS
problemSize = 1021 | A B^T | 3183 -> 3165 (-18) GFLOPS
problemSize = 1022 | A B^T | 3214 -> 3200 (-14) GFLOPS
problemSize = 1023 | A B^T | 3212 -> 3193 (-19) GFLOPS
problemSize = 1024 | A B^T | 3392 -> 3377 (-15) GFLOPS
problemSize = 1025 | A B^T | 3039 -> 3034 (-5) GFLOPS
problemSize = 1026 | A B^T | 3055 -> 3044 (-11) GFLOPS
problemSize = 1027 | A B^T | 3042 -> 3029 (-13) GFLOPS
problemSize = 1028 | A B^T | 3082 -> 3088 (+6) GFLOPS
problemSize = 1029 | A B^T | 3059 -> 3049 (-10) GFLOPS
problemSize = 1030 | A B^T | 3076 -> 3068 (-8) GFLOPS
problemSize = 1031 | A B^T | 3069 -> 3060 (-9) GFLOPS
problemSize = 1032 | A B^T | 3140 -> 3206 (+66) GFLOPS
problemSize = 1033 | A B^T | 2877 -> 2947 (+70) GFLOPS
problemSize = 1034 | A B^T | 2729 -> 2807 (+78) GFLOPS
problemSize = 1035 | A B^T | 2604 -> 2640 (+36) GFLOPS
problemSize = 1036 | A B^T | 2586 -> 2562 (-24) GFLOPS
problemSize = 1037 | A B^T | 2589 -> 2570 (-19) GFLOPS
problemSize = 1038 | A B^T | 2632 -> 2612 (-20) GFLOPS
problemSize = 1039 | A B^T | 2666 -> 2626 (-40) GFLOPS
problemSize = 1040 | A B^T | 2704 -> 2672 (-32) GFLOPS
problemSize = 1041 | A B^T | 2710 -> 2704 (-6) GFLOPS
problemSize = 1042 | A B^T | 2741 -> 2701 (-40) GFLOPS
problemSize = 1043 | A B^T | 2734 -> 2707 (-27) GFLOPS
problemSize = 1044 | A B^T | 2708 -> 2714 (+6) GFLOPS
problemSize = 1045 | A B^T | 2667 -> 2666 (-1) GFLOPS
problemSize = 1046 | A B^T | 2643 -> 2632 (-11) GFLOPS
problemSize = 1047 | A B^T | 2664 -> 2640 (-24) GFLOPS
problemSize = 1048 | A B^T | 2700 -> 2662 (-38) GFLOPS
problemSize = 1049 | A B^T | 2765 -> 2708 (-57) GFLOPS
problemSize = 1050 | A B^T | 2772 -> 2737 (-35) GFLOPS
problemSize = 1051 | A B^T | 2766 -> 2764 (-2) GFLOPS
problemSize = 1052 | A B^T | 2763 -> 2750 (-13) GFLOPS
problemSize = 1053 | A B^T | 2782 -> 2769 (-13) GFLOPS
problemSize = 1054 | A B^T | 2804 -> 2777 (-27) GFLOPS
problemSize = 1055 | A B^T | 2789 -> 2780 (-9) GFLOPS
problemSize = 1056 | A B^T | 2825 -> 2779 (-46) GFLOPS
problemSize = 1057 | A B^T | 2645 -> 2666 (+21) GFLOPS
problemSize = 1058 | A B^T | 2641 -> 2626 (-15) GFLOPS
problemSize = 1059 | A B^T | 2665 -> 2628 (-37) GFLOPS
problemSize = 1060 | A B^T | 2654 -> 2645 (-9) GFLOPS
problemSize = 1061 | A B^T | 2662 -> 2632 (-30) GFLOPS
problemSize = 1062 | A B^T | 2690 -> 2679 (-11) GFLOPS
problemSize = 1063 | A B^T | 2698 -> 2668 (-30) GFLOPS
problemSize = 1064 | A B^T | 2672 -> 2677 (+5) GFLOPS
problemSize = 1065 | A B^T | 2681 -> 2714 (+33) GFLOPS
problemSize = 1066 | A B^T | 2664 -> 2662 (-2) GFLOPS
problemSize = 1067 | A B^T | 2702 -> 2647 (-55) GFLOPS
problemSize = 1068 | A B^T | 2716 -> 2697 (-19) GFLOPS
problemSize = 1069 | A B^T | 2723 -> 2684 (-39) GFLOPS
problemSize = 1070 | A B^T | 2734 -> 2702 (-32) GFLOPS
problemSize = 1071 | A B^T | 2706 -> 2708 (+2) GFLOPS
problemSize = 1072 | A B^T | 2686 -> 2703 (+17) GFLOPS
M4, BF16
problemSize = 976 | A B | 3225 -> 3224 (-1) GFLOPS
problemSize = 977 | A B | 3198 -> 3213 (+15) GFLOPS
problemSize = 978 | A B | 3212 -> 3214 (+2) GFLOPS
problemSize = 979 | A B | 3219 -> 3220 (+1) GFLOPS
problemSize = 980 | A B | 3225 -> 3243 (+18) GFLOPS
problemSize = 981 | A B | 3243 -> 3240 (-3) GFLOPS
problemSize = 982 | A B | 3247 -> 3251 (+4) GFLOPS
problemSize = 983 | A B | 3254 -> 3266 (+12) GFLOPS
problemSize = 984 | A B | 3273 -> 3270 (-3) GFLOPS
problemSize = 985 | A B | 3252 -> 3254 (+2) GFLOPS
problemSize = 986 | A B | 3261 -> 3270 (+9) GFLOPS
problemSize = 987 | A B | 3275 -> 3269 (-6) GFLOPS
problemSize = 988 | A B | 3282 -> 3290 (+8) GFLOPS
problemSize = 989 | A B | 3290 -> 3294 (+4) GFLOPS
problemSize = 990 | A B | 3295 -> 3299 (+4) GFLOPS
problemSize = 991 | A B | 3311 -> 3311 ( 0) GFLOPS
problemSize = 992 | A B | 3325 -> 3319 (-6) GFLOPS
problemSize = 993 | A B | 3115 -> 3127 (+12) GFLOPS
problemSize = 994 | A B | 3118 -> 3132 (+14) GFLOPS
problemSize = 995 | A B | 3130 -> 3140 (+10) GFLOPS
problemSize = 996 | A B | 3141 -> 3154 (+13) GFLOPS
problemSize = 997 | A B | 3150 -> 3160 (+10) GFLOPS
problemSize = 998 | A B | 3164 -> 3169 (+5) GFLOPS
problemSize = 999 | A B | 3166 -> 3182 (+16) GFLOPS
problemSize = 1000 | A B | 3183 -> 3181 (-2) GFLOPS
problemSize = 1001 | A B | 3162 -> 3174 (+12) GFLOPS
problemSize = 1002 | A B | 3172 -> 3183 (+11) GFLOPS
problemSize = 1003 | A B | 3180 -> 3191 (+11) GFLOPS
problemSize = 1004 | A B | 3196 -> 3205 (+9) GFLOPS
problemSize = 1005 | A B | 3203 -> 3210 (+7) GFLOPS
problemSize = 1006 | A B | 3207 -> 3220 (+13) GFLOPS
problemSize = 1007 | A B | 3221 -> 3228 (+7) GFLOPS
problemSize = 1008 | A B | 3230 -> 3233 (+3) GFLOPS
problemSize = 1009 | A B | 3216 -> 3227 (+11) GFLOPS
problemSize = 1010 | A B | 3222 -> 3229 (+7) GFLOPS
problemSize = 1011 | A B | 3234 -> 3244 (+10) GFLOPS
problemSize = 1012 | A B | 3243 -> 3257 (+14) GFLOPS
problemSize = 1013 | A B | 3254 -> 3262 (+8) GFLOPS
problemSize = 1014 | A B | 3261 -> 3268 (+7) GFLOPS
problemSize = 1015 | A B | 3271 -> 3280 (+9) GFLOPS
problemSize = 1016 | A B | 3283 -> 3285 (+2) GFLOPS
problemSize = 1017 | A B | 3266 -> 3276 (+10) GFLOPS
problemSize = 1018 | A B | 3275 -> 3279 (+4) GFLOPS
problemSize = 1019 | A B | 3288 -> 3294 (+6) GFLOPS
problemSize = 1020 | A B | 3294 -> 3306 (+12) GFLOPS
problemSize = 1021 | A B | 3306 -> 3311 (+5) GFLOPS
problemSize = 1022 | A B | 3312 -> 3318 (+6) GFLOPS
problemSize = 1023 | A B | 3330 -> 3331 (+1) GFLOPS
problemSize = 1024 | A B | 3313 -> 3340 (+27) GFLOPS
problemSize = 1025 | A B | 3153 -> 3155 (+2) GFLOPS
problemSize = 1026 | A B | 3158 -> 3162 (+4) GFLOPS
problemSize = 1027 | A B | 3166 -> 3171 (+5) GFLOPS
problemSize = 1028 | A B | 3175 -> 3184 (+9) GFLOPS
problemSize = 1029 | A B | 3184 -> 3187 (+3) GFLOPS
problemSize = 1030 | A B | 3194 -> 3197 (+3) GFLOPS
problemSize = 1031 | A B | 3204 -> 3203 (-1) GFLOPS
problemSize = 1032 | A B | 3210 -> 3216 (+6) GFLOPS
problemSize = 1033 | A B | 3197 -> 3196 (-1) GFLOPS
problemSize = 1034 | A B | 3191 -> 3210 (+19) GFLOPS
problemSize = 1035 | A B | 3218 -> 3220 (+2) GFLOPS
problemSize = 1036 | A B | 3223 -> 3238 (+15) GFLOPS
problemSize = 1037 | A B | 3236 -> 3237 (+1) GFLOPS
problemSize = 1038 | A B | 3238 -> 3246 (+8) GFLOPS
problemSize = 1039 | A B | 3251 -> 3253 (+2) GFLOPS
problemSize = 1040 | A B | 3267 -> 3268 (+1) GFLOPS
problemSize = 1041 | A B | 3250 -> 3249 (-1) GFLOPS
problemSize = 1042 | A B | 3255 -> 3260 (+5) GFLOPS
problemSize = 1043 | A B | 3268 -> 3270 (+2) GFLOPS
problemSize = 1044 | A B | 3270 -> 3279 (+9) GFLOPS
problemSize = 1045 | A B | 3284 -> 3284 ( 0) GFLOPS
problemSize = 1046 | A B | 3288 -> 3295 (+7) GFLOPS
problemSize = 1047 | A B | 3302 -> 3302 ( 0) GFLOPS
problemSize = 1048 | A B | 3316 -> 3314 (-2) GFLOPS
problemSize = 1049 | A B | 3297 -> 3299 (+2) GFLOPS
problemSize = 1050 | A B | 3304 -> 3308 (+4) GFLOPS
problemSize = 1051 | A B | 3316 -> 3310 (-6) GFLOPS
problemSize = 1052 | A B | 3326 -> 3335 (+9) GFLOPS
problemSize = 1053 | A B | 3337 -> 3336 (-1) GFLOPS
problemSize = 1054 | A B | 3343 -> 3342 (-1) GFLOPS
problemSize = 1055 | A B | 3355 -> 3349 (-6) GFLOPS
problemSize = 1056 | A B | 3371 -> 3370 (-1) GFLOPS
problemSize = 1057 | A B | 3180 -> 3194 (+14) GFLOPS
problemSize = 1058 | A B | 3188 -> 3196 (+8) GFLOPS
problemSize = 1059 | A B | 3197 -> 3201 (+4) GFLOPS
problemSize = 1060 | A B | 3203 -> 3221 (+18) GFLOPS
problemSize = 1061 | A B | 3218 -> 3221 (+3) GFLOPS
problemSize = 1062 | A B | 2947 -> 3073 (+126) GFLOPS
problemSize = 1063 | A B | 2865 -> 2914 (+49) GFLOPS
problemSize = 1064 | A B | 2800 -> 2853 (+53) GFLOPS
problemSize = 1065 | A B | 2784 -> 2833 (+49) GFLOPS
problemSize = 1066 | A B | 2798 -> 2794 (-4) GFLOPS
problemSize = 1067 | A B | 2810 -> 2807 (-3) GFLOPS
problemSize = 1068 | A B | 2818 -> 2830 (+12) GFLOPS
problemSize = 1069 | A B | 2845 -> 2860 (+15) GFLOPS
problemSize = 1070 | A B | 2864 -> 2867 (+3) GFLOPS
problemSize = 1071 | A B | 2881 -> 2885 (+4) GFLOPS
problemSize = 1072 | A B | 2866 -> 2878 (+12) GFLOPS
problemSize = 976 | A B^T | 3196 -> 3199 (+3) GFLOPS
problemSize = 977 | A B^T | 3168 -> 3156 (-12) GFLOPS
problemSize = 978 | A B^T | 3177 -> 3156 (-21) GFLOPS
problemSize = 979 | A B^T | 3182 -> 3166 (-16) GFLOPS
problemSize = 980 | A B^T | 3193 -> 3181 (-12) GFLOPS
problemSize = 981 | A B^T | 3209 -> 3180 (-29) GFLOPS
problemSize = 982 | A B^T | 3206 -> 3193 (-13) GFLOPS
problemSize = 983 | A B^T | 3230 -> 3206 (-24) GFLOPS
problemSize = 984 | A B^T | 3241 -> 3237 (-4) GFLOPS
problemSize = 985 | A B^T | 3219 -> 3195 (-24) GFLOPS
problemSize = 986 | A B^T | 3226 -> 3206 (-20) GFLOPS
problemSize = 987 | A B^T | 3240 -> 3210 (-30) GFLOPS
problemSize = 988 | A B^T | 3245 -> 3233 (-12) GFLOPS
problemSize = 989 | A B^T | 3252 -> 3232 (-20) GFLOPS
problemSize = 990 | A B^T | 3260 -> 3240 (-20) GFLOPS
problemSize = 991 | A B^T | 3271 -> 3251 (-20) GFLOPS
problemSize = 992 | A B^T | 3312 -> 3288 (-24) GFLOPS
problemSize = 993 | A B^T | 3068 -> 3062 (-6) GFLOPS
problemSize = 994 | A B^T | 3079 -> 3075 (-4) GFLOPS
problemSize = 995 | A B^T | 3086 -> 3081 (-5) GFLOPS
problemSize = 996 | A B^T | 3100 -> 3094 (-6) GFLOPS
problemSize = 997 | A B^T | 3103 -> 3091 (-12) GFLOPS
problemSize = 998 | A B^T | 3115 -> 3107 (-8) GFLOPS
problemSize = 999 | A B^T | 3124 -> 3118 (-6) GFLOPS
problemSize = 1000 | A B^T | 3145 -> 3142 (-3) GFLOPS
problemSize = 1001 | A B^T | 3119 -> 3113 (-6) GFLOPS
problemSize = 1002 | A B^T | 3131 -> 3124 (-7) GFLOPS
problemSize = 1003 | A B^T | 3138 -> 3129 (-9) GFLOPS
problemSize = 1004 | A B^T | 3153 -> 3144 (-9) GFLOPS
problemSize = 1005 | A B^T | 3156 -> 3145 (-11) GFLOPS
problemSize = 1006 | A B^T | 3167 -> 3160 (-7) GFLOPS
problemSize = 1007 | A B^T | 3182 -> 3164 (-18) GFLOPS
problemSize = 1008 | A B^T | 3206 -> 3206 ( 0) GFLOPS
problemSize = 1009 | A B^T | 3173 -> 3159 (-14) GFLOPS
problemSize = 1010 | A B^T | 3183 -> 3171 (-12) GFLOPS
problemSize = 1011 | A B^T | 3188 -> 3176 (-12) GFLOPS
problemSize = 1012 | A B^T | 3201 -> 3194 (-7) GFLOPS
problemSize = 1013 | A B^T | 3204 -> 3192 (-12) GFLOPS
problemSize = 1014 | A B^T | 3216 -> 3202 (-14) GFLOPS
problemSize = 1015 | A B^T | 3217 -> 3203 (-14) GFLOPS
problemSize = 1016 | A B^T | 3247 -> 3244 (-3) GFLOPS
problemSize = 1017 | A B^T | 3215 -> 3204 (-11) GFLOPS
problemSize = 1018 | A B^T | 3231 -> 3215 (-16) GFLOPS
problemSize = 1019 | A B^T | 3233 -> 3224 (-9) GFLOPS
problemSize = 1020 | A B^T | 3250 -> 3242 (-8) GFLOPS
problemSize = 1021 | A B^T | 3252 -> 3234 (-18) GFLOPS
problemSize = 1022 | A B^T | 3266 -> 3249 (-17) GFLOPS
problemSize = 1023 | A B^T | 3272 -> 3255 (-17) GFLOPS
problemSize = 1024 | A B^T | 3322 -> 3303 (-19) GFLOPS
problemSize = 1025 | A B^T | 3096 -> 3080 (-16) GFLOPS
problemSize = 1026 | A B^T | 3111 -> 3091 (-20) GFLOPS
problemSize = 1027 | A B^T | 3108 -> 3091 (-17) GFLOPS
problemSize = 1028 | A B^T | 3129 -> 3118 (-11) GFLOPS
problemSize = 1029 | A B^T | 3127 -> 3116 (-11) GFLOPS
problemSize = 1030 | A B^T | 3146 -> 3127 (-19) GFLOPS
problemSize = 1031 | A B^T | 3139 -> 3131 (-8) GFLOPS
problemSize = 1032 | A B^T | 3169 -> 3160 (-9) GFLOPS
problemSize = 1033 | A B^T | 3142 -> 3120 (-22) GFLOPS
problemSize = 1034 | A B^T | 3157 -> 3139 (-18) GFLOPS
problemSize = 1035 | A B^T | 3162 -> 3145 (-17) GFLOPS
problemSize = 1036 | A B^T | 3181 -> 3166 (-15) GFLOPS
problemSize = 1037 | A B^T | 3179 -> 3164 (-15) GFLOPS
problemSize = 1038 | A B^T | 3196 -> 3179 (-17) GFLOPS
problemSize = 1039 | A B^T | 3198 -> 3185 (-13) GFLOPS
problemSize = 1040 | A B^T | 3231 -> 3224 (-7) GFLOPS
problemSize = 1041 | A B^T | 3195 -> 3178 (-17) GFLOPS
problemSize = 1042 | A B^T | 3211 -> 3190 (-21) GFLOPS
problemSize = 1043 | A B^T | 3212 -> 3193 (-19) GFLOPS
problemSize = 1044 | A B^T | 3234 -> 3215 (-19) GFLOPS
problemSize = 1045 | A B^T | 3226 -> 3210 (-16) GFLOPS
problemSize = 1046 | A B^T | 3246 -> 3227 (-19) GFLOPS
problemSize = 1047 | A B^T | 3249 -> 3230 (-19) GFLOPS
problemSize = 1048 | A B^T | 3270 -> 3263 (-7) GFLOPS
problemSize = 1049 | A B^T | 3242 -> 3224 (-18) GFLOPS
problemSize = 1050 | A B^T | 3046 -> 3096 (+50) GFLOPS
problemSize = 1051 | A B^T | 2928 -> 2964 (+36) GFLOPS
problemSize = 1052 | A B^T | 2882 -> 2887 (+5) GFLOPS
problemSize = 1053 | A B^T | 2839 -> 2853 (+14) GFLOPS
problemSize = 1054 | A B^T | 2831 -> 2801 (-30) GFLOPS
problemSize = 1055 | A B^T | 2815 -> 2819 (+4) GFLOPS
problemSize = 1056 | A B^T | 2923 -> 2868 (-55) GFLOPS
problemSize = 1057 | A B^T | 2768 -> 2791 (+23) GFLOPS
problemSize = 1058 | A B^T | 2747 -> 2735 (-12) GFLOPS
problemSize = 1059 | A B^T | 2723 -> 2711 (-12) GFLOPS
problemSize = 1060 | A B^T | 2758 -> 2729 (-29) GFLOPS
problemSize = 1061 | A B^T | 2741 -> 2743 (+2) GFLOPS
problemSize = 1062 | A B^T | 2742 -> 2736 (-6) GFLOPS
problemSize = 1063 | A B^T | 2746 -> 2719 (-27) GFLOPS
problemSize = 1064 | A B^T | 2764 -> 2752 (-12) GFLOPS
problemSize = 1065 | A B^T | 2799 -> 2803 (+4) GFLOPS
problemSize = 1066 | A B^T | 2742 -> 2756 (+14) GFLOPS
problemSize = 1067 | A B^T | 2735 -> 2737 (+2) GFLOPS
problemSize = 1068 | A B^T | 2775 -> 2742 (-33) GFLOPS
problemSize = 1069 | A B^T | 2774 -> 2746 (-28) GFLOPS
problemSize = 1070 | A B^T | 2812 -> 2771 (-41) GFLOPS
problemSize = 1071 | A B^T | 2784 -> 2783 (-1) GFLOPS
problemSize = 1072 | A B^T | 2806 -> 2794 (-12) GFLOPS
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment