Welcome to
mirror list
, hosted at
ThFree Co
, Russian Federation.
github.com/marian-nmt/FBGEMM.git - Unnamed repository; edit this file 'description' to name the repository.
index
:
github.com/marian-nmt/FBGEMM.git
aaronpburke/fix-install-targets2
copyPublic
gcc11support
master
youki/avx512-avx2
youki/benchmarksparse
youki/fix-avx2-fp16
youki/fix-gcc10-compile
youki/fix-gcc9-build
youki/fix-stdexcept
youki/fp16avx512
youki/fp16intrinsic
youki/improve-mem-alloc
youki/improve-mem-alloc-marian
youki/jit-experiments
youki/merge-win-int8
youki/mergemaster01092020
youki/mergemaster1206
youki/static-code-cache
youki/testsparse
youki/unit-test-fix
youki/unordered_map
youki/upstream0217
youki/upstream0509
Unnamed repository; edit this file 'description' to name the repository.
www-data
summary
refs
log
tree
commit
diff
log msg
author
committer
range
path:
root
/
src
Age
Commit message (
Expand
)
Author
2019-07-16
unpack through unified convolution interface (#105)
Daya Khudia
2019-07-16
Assume input weights to be in transposed format for convUnified (#104)
Daya Khudia
2019-07-10
Refactoring unpack weight function (#103)
Jianyu Huang
2019-07-06
Unpack data for 3x3 (and 3x3x3) depthwise convolution
Daya Khudia
2019-07-06
Implement ::unpack() for PackWeightMatrixForGConv
Jaewon Lee
2019-07-01
Refactor the code and avoid the duplication (#102)
Jianyu Huang
2019-07-01
Clean up some code for JIT code generator (#101)
Jianyu Huang
2019-06-20
Per channel and groupwise quantization (#99)
Daya Khudia
2019-06-15
Update the logic of checking valid parameters.
Mike Tsai
2019-06-12
Print packed matrix for each group as well
Daya Khudia
2019-06-07
Remove duplicated header and undo some changes in D15399811
Daya Khudia
2019-06-05
Unified convolution interface
Daya Khudia
2019-06-04
Add quantized::fbgemm_linear_unpack operator for serialization (#97)
Jianyu Huang
2019-05-24
Fix kernel logging
Mike Tsai
2019-05-16
fixing compiler warnings for uninitialized MR, NCB, KCB
Protonu Basu
2019-04-19
make sure cpuinfo_initialize called before fbgemmHasAvx2/512Support (#94)
Jongsoo Park
2019-04-03
optimize dw conv for symmetric quant (#73)
Jongsoo Park
2019-04-02
Exposing tuning parameters in FBGEMM (MCB, NCB, KCB, MR, NR, Row Interleave) ...
Protonu Basu
2019-03-25
Packing B documentation
Daya S Khudia
2019-03-21
Improves small N cases back to what they were
Daya S Khudia
2019-03-21
Allocate some registers for B matrix loading and reuse loaded results
Daya S Khudia
2019-03-21
Further optimize acc16 kernel and cache blocking dimension for B matrix is no...
Daya S Khudia
2019-03-21
Further optimize acc32 kernel and cache blocking dimension for B matrix is no...
Daya S Khudia
2019-03-19
Dump generated kernels in files
Daya S Khudia
2019-03-18
Add the Naive bfloat16 implementation based on MKL
Jianyu Huang
2019-03-13
optimize requantize for float out processing (#85)
Jongsoo Park
2019-03-08
No need for PackA when m==1 (#83)
Jianyu Huang
2019-03-08
Fixes for FBGEMM FP16 performance (#82)
Jianyu Huang
2019-03-06
Add Avx512BW/VL/DQ check (#84)
Jianyu Huang
2019-02-26
barebone int8-acc16 and int8-acc32 benchmarks
Daya S Khudia
2019-02-23
specialization for first conv (#80)
Jongsoo Park
2019-02-22
Optimize PackB routine by removing addr function
Jianyu Huang
2019-02-20
optimize PackAWithIm2Col for symmetric b quant
Jongsoo Park
2019-02-20
increase test coverage (#78)
Jongsoo Park
2019-02-19
remove unused member var kBlock_ (#77)
Jongsoo Park
2019-02-15
simple spmdm optimization (#76)
Jongsoo Park
2019-02-14
clean up depthwise conv interface (#72)
Jongsoo Park
2019-02-14
fix bug in group conv + avx512 (#75)
Jongsoo Park
2019-02-14
JIT kernel should only handle a small portion of NCB for the last block: mult...
Jianyu Huang
2019-02-14
Fix PackBMatrix<T, accT>::printPackedMatrix issues
Jianyu Huang
2019-02-13
optimize gconv for b symmetric quantization (#70)
Jongsoo Park
2019-02-13
no need to subtract col offset if a_zp is 0 (#69)
Jongsoo Park
2019-02-13
isZeroPointZero_ -> isAZeroPointZero_ (#71)
Jongsoo Park
2019-02-13
group conv optimized for 16 channels per group (#68)
Jongsoo Park
2019-02-02
gconv optimized for 8 channels per group (#65)
Jongsoo Park
2019-02-02
minor optimization in handling zero points for row offset (#63)
Jongsoo Park
2019-02-01
make G slowest moving dim of packed weight of gconv (#62)
Jongsoo Park
2019-02-01
more careful about movaps (#60)
Jongsoo Park
2019-02-01
specialized requantization for gconv (#61)
Jongsoo Park
2019-01-31
optimize requantization remainder (#64)
Jongsoo Park
[next]