Skip to content

Instantly share code, notes, and snippets.

Keno Fischer Keno

Block or report user

Report or block Keno

Hide content and notifications from this user.

Learn more about blocking users

Contact Support about this user’s behavior.

Learn more about reporting abuse

Report abuse
View GitHub Profile
View gist:0bf23c7d0c75cfdab84580039d54997e
module Branchy
using Base: significand_mask, significand_bits, exponent_mask, sign_mask
@inline function compute_base_shift(e)
if e < -25 # Very small numbers map to zero
base = 0x0000
sh = 25
elseif e < -14 # Small numbers map to denorms
base = 0x0000
sh = -e-1
View gist:5d168c39085312f6b7398555cdc07127
00:00.0 Host bridge [0600]: Intel Corporation Xeon E3-1200 v5/E3-1500 v5/6th Gen Core Processor Host Bridge/DRAM Registers [8086:190f] (rev 07)
Subsystem: Super Micro Computer Inc Xeon E3-1200 v5/E3-1500 v5/6th Gen Core Processor Host Bridge/DRAM Registers [15d9:0884]
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort+ >SERR- <PERR- INTx-
Latency: 0
Capabilities: [e0] Vendor Specific Information: Len=10 <?>
Kernel driver in use: skl_uncore
00: 86 80 0f 19 06 00 90 20 07 00 00 06 00 00 00 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 d9 15 84 08
View gist:4751ff94f10868275b888f957aef0c64
/workspace/srcdir/julia/usr/share/julia/stdlib/v1.3/PkInternal error: encountered unexpected error in runtime:
TypeError(func=:Tuple, context="parameter", expected=Type, got=<?#0x7f27f22266b0::<circular reference @-1>>)
rec_backtrace at /workspace/srcdir/julia/src/stackwalk.c:94
record_backtrace at /workspace/srcdir/julia/src/task.c:216 [inlined]
jl_throw at /workspace/srcdir/julia/src/task.c:442
jl_type_error_rt at /workspace/srcdir/julia/src/rtutils.c:118
jl_f_apply_type at /workspace/srcdir/julia/src/builtins.c:927
jl_apply_generic at /workspace/srcdir/julia/src/gf.c:2192
jl_apply at /workspace/srcdir/julia/src/julia.h:1609 [inlined]
jl_f__apply at /workspace/srcdir/julia/src/builtins.c:626
View scev_umin_llvm80.patch
commit 18e563f695dd561c32393512fbdb8ce8771d7e5f
Author: Keno Fischer <keno@juliacomputing.com>
Date: Thu May 2 08:35:22 2019 -0400
[SCEV] Add explicit representations of umin/smin
Summary:
Currently we express umin as `~umax(~x, ~y)`. However, this becomes
a problem for operands in non-integral pointer spaces, because `~x`
is not something we can compute for `x` non-integral. However, since
View serialdump
DRAM Init-DDR3
CBR0-1357135670123456701234567
CBR134Done
Loading kernel: [4038209a]
Loading device tree: [35cbc4d5]
Loading initrd: [087d899b]
Starting kernel ...
Uncompressing Linux... done, booting the kernel.
View gist:b918b17e80d06760679c65b9443806db
2 1 ─ %1 = Main.randn(Main.Float32, 20)::Array{Float32,1} │
3 │ %2 = Zygote.unwrap(%1)::Array{Float32,1} │╻╷╷╷╷ gradient
│ %3 = Zygote.unwrap(_3)::Float32 ││╻╷ #21
│ %4 = Zygote.unwrap(_2)::Chain{Tuple{Dense{typeof(identity),Array{Float32,2},Array{Float32,1}}}} │││╻╷╷╷╷╷╷ gradient
│ %5 = Zygote.unwrap(%2)::Array{Float32,1} ││││╻╷╷╷╷ #22
│ %6 = Core.tuple(%5, %3)::Tuple{Array{Float32,1},Float32} │││││╻ _forward
│ %7 = Zygote.compute_szs(%6, 1)::Array{Int64,1}
View gist:3bba34b85f80d37c59a9ffcb39c06439
HloModule test
comp2 {
ROOT comp2_constant3012 = f32[] constant(0)
}
comp3 {
ROOT comp3_constant3013 = f32[] constant(1)
}
View gist:68694aaf47b2c7e0a4561ec0a23e6f2c
HloModule test
comp1 {
comp1_parameter3867 = f32[] parameter(0)
comp1_parameter3868 = f32[] parameter(1)
ROOT comp1_divide3869 = f32[] divide(comp1_parameter3867, comp1_parameter3868)
}
comp2 {
comp2_parameter3875 = f32[] parameter(0)
View gist:4b290b1a43ae06e7a69b162a9ca9080d
HloModule test
comp1 {
comp1_parameter3990 = f32[] parameter(0)
comp1_parameter3991 = f32[] parameter(1)
ROOT comp1_divide3992 = f32[] divide(comp1_parameter3990, comp1_parameter3991)
}
comp2 {
comp2_parameter3998 = f32[] parameter(0)
View gist:e6325f1f9dc3825630019def9ce8bf01
CodeInfo(
41 1 ─ %1 = Base.sin::typeof(sin) │╻╷╷╷╷╷╷ derivative
│ %2 = invoke %1(_2::Float64)::Float64 ││┃│││││ gradient
│ %3 = Base.cos::typeof(cos) │││╻ forward
│ %4 = invoke %3(_2::Float64)::Float64 ││││┃│││ _forward
│ %5 = Base.exp::typeof(exp) │││││╻╷ fx_opt
│ %6 = invoke %5(_2::Float64)::Float64
You can’t perform that action at this time.