sbahra

This isn't exactly a formal performance comparison. First column is length of
buffer in bytes, second column is the libc implementation and third column is
my implementation.

nehalem:string sbahra$ uname -a
Darwin nehalem.home 10.6.0 Darwin Kernel Version 10.6.0: Wed Nov 10 18:13:17 PST 2010; root:xnu-1504.9.26~3/RELEASE_I386 i386
nehalem:string sbahra$ for j in memchr memset strchr strlen; do echo $j; for i in 32 64 128 256 512 1024; do ./${j} $i; done; done
memchr
           32              214               61
           64              279               65

sbahra

Tommy benchmark program.
      100000     concurrencykit    forward
   forward,     insert, concurrencykit,   82 [ns]
   forward,     change, concurrencykit,  169 [ns]
   forward,        hit, concurrencykit,   84 [ns]
   forward,       miss, concurrencykit,   79 [ns]
   forward,     remove, concurrencykit,   80 [ns]
      100000     concurrencykit     random
    random,     insert, concurrencykit,   82 [ns]
    random,     change, concurrencykit,  175 [ns]

sbahra

Tommy benchmark program.
       10000     concurrencykit    forward
   forward,     insert, concurrencykit,   88 [ns]
   forward,     change, concurrencykit,  112 [ns]
   forward,        hit, concurrencykit,   46 [ns]
   forward,       miss, concurrencykit,   62 [ns]
   forward,     remove, concurrencykit,   61 [ns]
       10000     concurrencykit     random
    random,     insert, concurrencykit,   83 [ns]
    random,     change, concurrencykit,  111 [ns]

sbahra

  sbahra@broadwell:~/projects/tommyds/benchmark$ LD_PRELOAD=/opt/backtrace/lib/libck.so ../benchmark.test -n 100000 -d concurrencykit
Tommy benchmark program.
      100000     concurrencykit    forward
   forward,     insert, concurrencykit,   99 [ns]
   forward,     change, concurrencykit,   91 [ns]
   forward,        hit, concurrencykit,   83 [ns]
   forward,       miss, concurrencykit,  101 [ns]
   forward,     remove, concurrencykit,   45 [ns]
      100000     concurrencykit     random
    random,     insert, concurrencykit,   94 [ns]

sbahra

haswell:benchmark sbahra$ ./latency
WRITE: brlock                33
WRITE: naive                 31
READ:  brlock                38
READ:  naive                 35
haswell:benchmark sbahra$ ./throughput 1 4
Creating threads (brlock)...done
Waiting for threads to finish acquisition regression...done

         1                   51

sbahra

Keybase proof

I hereby claim:

• I am sbahra on github.
• I am sbahra (https://keybase.io/sbahra) on keybase.
• I have a public key whose fingerprint is 8EE0 B9E3 185A AA72 99E6 0BE5 19D2 E15A 9105 9E27

To claim this, I am signing this object:

sbahra

### Keybase proof

I hereby claim:

  * I am sbahra on github.
  * I am sbahra (https://keybase.io/sbahra) on keybase.
  * I have a public key whose fingerprint is 8EE0 B9E3 185A AA72 99E6  0BE5 19D2 E15A 9105 9E27

To claim this, I am signing this object:

sbahra

PID: 21634
--------------------------------------------------------------------------------
Thread 21640
  [  0] libpthread-2.13.so       pause (../sysdeps/unix/syscall-template.S:82)
        => mov (%rsp), %rdi
  [  1] crash                    crash_signal (crash.c:511)
        => call 0xffffffffffffd0ae
     s = (parameter) 11
     si = (parameter) reference(0x7fb8cdbfa9e0, 0x7fb8cdbfab30)
       -><> = structure

sbahra

diff --git a/configure.ac b/configure.ac
index db006a8..1112e73 100644
--- a/configure.ac
+++ b/configure.ac
@@ -42,6 +42,19 @@ AC_ARG_ENABLE(lenient, [
    lenient=$enableval
 ])
 
+rtm=no
+AC_ARG_ENABLE(rtm, [

sbahra

i5:ck sbahra$ git diff
diff --git a/include/ck_epoch.h b/include/ck_epoch.h
index 1b7d1b5..dd2f998 100644
--- a/include/ck_epoch.h
+++ b/include/ck_epoch.h
@@ -147,7 +147,7 @@ ck_epoch_call(ck_epoch_t *epoch,
 void ck_epoch_init(ck_epoch_t *);
 ck_epoch_record_t *ck_epoch_recycle(ck_epoch_t *);
 void ck_epoch_register(ck_epoch_t *, ck_epoch_record_t *);
-void ck_epoch_unregister(ck_epoch_record_t *);