Youmu Liu
|
413e41db45
|
other gemm kernel use half-vector-machine style
|
2026-06-08 18:09:15 +08:00 |
|
Youmu Liu
|
9d5c0fd8b4
|
other gemm kernel use half-vector-machine style
|
2026-06-08 16:18:03 +08:00 |
|
Nitload
|
1c515ea2ae
|
FMA-avx512-f64 kernel use half-vector-machine design
|
2026-06-08 01:02:34 +08:00 |
|
Youmu Liu
|
4336b0ca5a
|
Move some marcros from x86 ro base head
|
2026-06-05 13:32:59 +08:00 |
|
Youmu Liu
|
94b2d759ae
|
Gemm/Fma kernel on avx2 use new solution, with less permute and shuff instruction
|
2026-06-04 21:05:38 +08:00 |
|
Youmu Liu
|
2690cf9341
|
Gemm/Fma kernel on avx512 use new solution, with less permute and shuff instruction
|
2026-06-04 16:43:17 +08:00 |
|
Youmu Liu
|
6c38bcfade
|
Find solution to fma, change some type define macros
|
2026-06-03 18:06:37 +08:00 |
|
Youmu Liu
|
9f3fc3b345
|
Compress L4 metadata into 32 bytes
|
2026-06-02 18:27:58 +08:00 |
|
Youmu Liu
|
ddb59c38ce
|
Little change in C code, remove isa specific code in std header
|
2026-05-28 18:44:53 +08:00 |
|
Youmu Liu
|
a99b19fa3f
|
add libs' cmakelists file
|
2026-05-11 18:11:28 +08:00 |
|
Youmu Liu
|
5706c89563
|
Remove ignored files
|
2026-05-11 09:30:04 +08:00 |
|
Nitload
|
156f6afe85
|
after merge
|
2026-05-04 17:25:01 +08:00 |
|
Youmu Liu
|
ae664cb541
|
matrix parallel start to use two different policy
|
2026-04-27 20:08:10 +08:00 |
|
Youmu Liu
|
ecb1f04770
|
matrix thread pool use ParaWaveFront and heavy sync
|
2026-04-26 17:38:10 +08:00 |
|
Youmu Liu
|
251762791e
|
L4 will not main stream in matrix schedualing
|
2026-04-24 17:31:49 +08:00 |
|
Youmu Liu
|
41350b0487
|
add a pre-kP evaluation function in c
|
2026-04-24 15:46:58 +08:00 |
|
Youmu Liu
|
2203a46eb7
|
can create and retire matrix thread by numa infomation
|
2026-04-15 18:08:08 +08:00 |
|
Youmu Liu
|
fdf11c729f
|
find a way to create Thread binded to LogicalCore properly
|
2026-04-14 18:08:22 +08:00 |
|
Youmu Liu
|
8a66c1697a
|
L4 completely add support for multi predictor, working on ver 5 mat para
|
2026-04-10 18:05:40 +08:00 |
|
Nitload
|
9fb08b672d
|
finally find pos to inject prediction to predictor
|
2026-04-08 00:25:21 +08:00 |
|
Youmu Liu
|
94a3127c0d
|
small changed
|
2026-04-07 18:08:34 +08:00 |
|
Nitload
|
12a216d404
|
new compute kernel design, less mem read delay
|
2026-04-07 01:04:05 +08:00 |
|
Youmu Liu
|
d702b3224c
|
adjust for clang-gcc
|
2026-04-03 18:40:11 +08:00 |
|
Youmu Liu
|
2f74e5e8a9
|
add startup numa sense
|
2026-04-02 18:13:32 +08:00 |
|
Youmu Liu
|
61303635ad
|
mat queue find solution, native supports clang
|
2026-03-27 18:14:54 +08:00 |
|
Youmu Liu
|
06e3aa119b
|
mat
|
2026-03-26 18:02:13 +08:00 |
|
Youmu Liu
|
2149bbeede
|
Normalize all existing files to LF
|
2026-03-06 18:31:11 +08:00 |
|
Youmu Liu
|
83997f64d0
|
Better FMA
|
2026-03-05 18:04:34 +08:00 |
|
Youmu Liu
|
dfa6f4ba92
|
Better L4 and array storage design
|
2026-03-04 18:12:58 +08:00 |
|
Nitload
|
7e970352ab
|
Better module loading.
|
2026-03-04 00:51:29 +08:00 |
|
Youmu Liu
|
91e136f81f
|
Working on L4, now on swiss. FMA_TILE is re arranged.
|
2026-03-03 18:02:31 +08:00 |
|
Youmu Liu
|
7e3ed8b0fd
|
decide to use L4 to solve numa and pack problem for large compute.
|
2026-03-02 17:33:03 +08:00 |
|
Nitload
|
67fbbde449
|
mat proc kernel std tamplate is ok, merging old decl
|
2026-03-02 00:24:15 +08:00 |
|
Youmu Liu
|
6e07e67835
|
new mat proc std
|
2026-02-28 18:08:26 +08:00 |
|
Youmu Liu
|
a798ab07f1
|
arrange of proj structure
|
2026-02-27 17:09:13 +08:00 |
|
Youmu Liu
|
15434ed053
|
many things
|
2026-02-26 18:05:35 +08:00 |
|
Youmu Liu
|
1b04196314
|
many versions of simdid
|
2026-02-25 18:02:56 +08:00 |
|
Nitload
|
b71515b2d4
|
per thread context loading is ok, working on global context
|
2026-02-25 00:20:09 +08:00 |
|
Youmu Liu
|
7ee9c749e5
|
working on per-PoolThread context loading and global context loading
|
2026-02-24 18:05:16 +08:00 |
|
Youmu Liu
|
d62e59cb93
|
Switch matrix parallel to task stealing, type and wrap are refacored, thread execution core has not yet
|
2026-02-24 15:12:23 +08:00 |
|
Nitload
|
d99c1e3aa6
|
add unicode figure
|
2026-02-23 19:54:21 +08:00 |
|
Nitload
|
c155cab510
|
HLSF use long as aloloc size
|
2026-02-23 12:32:15 +08:00 |
|
Youmu Liu
|
01984dee28
|
get the benchmark result, very POWERFULL
|
2026-01-29 20:05:53 +08:00 |
|
Youmu Liu
|
3881078770
|
add state nodes for regex state machine
|
2026-01-27 18:40:44 +08:00 |
|
Nitload
|
22baa6c5b0
|
Daily update on 20260125-013900 by Nitload Youmu. Additional Message : start working on state machine and regular analysis
Signed-off-by: Nitload <YoumuLiu_NSI@whv0t.onmicrosoft.com>
|
2026-01-25 01:39:00 +08:00 |
|
Youmu Liu
|
cb62bd48f3
|
new LU solvability for kernel,on dev, quality is briefly logic clear
|
2026-01-21 18:29:11 +08:00 |
|
Youmu Liu
|
cd36d0962b
|
new LU solvability for kernel,on dev
|
2026-01-20 13:52:02 +08:00 |
|
Youmu Liu
|
b7697d0342
|
add some mem opt struct from C# to C
|
2026-01-19 18:40:33 +08:00 |
|
Youmu Liu
|
11d544aed1
|
HLSf can now set max block size, not only 16MB
|
2026-01-16 17:52:58 +08:00 |
|
Youmu Liu
|
f02e55609b
|
solve or inv starts
|
2026-01-14 18:01:03 +08:00 |
|