Skip to content

Use intrinsics for all sifive_x280 kernels #822

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
devinamatthews merged 6 commits into from
Nov 29, 2024
Merged

Use intrinsics for all sifive_x280 kernels #822

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
1415e94
Use intrinsics for all sifive_x280 level 1, 1f, 1m kernels
myeh01 May 9, 2024
436ec9d
Reorder intrinsics
myeh01 Aug 1, 2024
f16ad58
Update toolchain
Aug 1, 2024
4e5e521
Update toolchain
Aug 5, 2024
b439815
Fix scal2v hack
myeh01 Aug 5, 2024
d167216
Use intrinsics for level 3 kernels
Sep 10, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
80 changes: 40 additions & 40 deletions config/sifive_x280/bli_cntx_init_sifive_x280.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -54,10 +54,10 @@ void bli_cntx_init_sifive_x280( cntx_t* cntx )
BLIS_ADDV_KER, BLIS_SCOMPLEX, bli_caddv_sifive_x280_intr,
BLIS_ADDV_KER, BLIS_DCOMPLEX, bli_zaddv_sifive_x280_intr,

BLIS_AMAXV_KER, BLIS_FLOAT, bli_samaxv_sifive_x280_asm,
BLIS_AMAXV_KER, BLIS_DOUBLE, bli_damaxv_sifive_x280_asm,
BLIS_AMAXV_KER, BLIS_SCOMPLEX, bli_camaxv_sifive_x280_asm,
BLIS_AMAXV_KER, BLIS_DCOMPLEX, bli_zamaxv_sifive_x280_asm,
BLIS_AMAXV_KER, BLIS_FLOAT, bli_samaxv_sifive_x280_intr,
BLIS_AMAXV_KER, BLIS_DOUBLE, bli_damaxv_sifive_x280_intr,
BLIS_AMAXV_KER, BLIS_SCOMPLEX, bli_camaxv_sifive_x280_intr,
BLIS_AMAXV_KER, BLIS_DCOMPLEX, bli_zamaxv_sifive_x280_intr,

BLIS_AXPBYV_KER, BLIS_FLOAT, bli_saxpbyv_sifive_x280_intr,
BLIS_AXPBYV_KER, BLIS_DOUBLE, bli_daxpbyv_sifive_x280_intr,
Expand All @@ -69,10 +69,10 @@ void bli_cntx_init_sifive_x280( cntx_t* cntx )
BLIS_AXPYV_KER, BLIS_SCOMPLEX, bli_caxpyv_sifive_x280_intr,
BLIS_AXPYV_KER, BLIS_DCOMPLEX, bli_zaxpyv_sifive_x280_intr,

BLIS_COPYV_KER, BLIS_FLOAT, bli_scopyv_sifive_x280_asm,
BLIS_COPYV_KER, BLIS_DOUBLE, bli_dcopyv_sifive_x280_asm,
BLIS_COPYV_KER, BLIS_SCOMPLEX, bli_ccopyv_sifive_x280_asm,
BLIS_COPYV_KER, BLIS_DCOMPLEX, bli_zcopyv_sifive_x280_asm,
BLIS_COPYV_KER, BLIS_FLOAT, bli_scopyv_sifive_x280_intr,
BLIS_COPYV_KER, BLIS_DOUBLE, bli_dcopyv_sifive_x280_intr,
BLIS_COPYV_KER, BLIS_SCOMPLEX, bli_ccopyv_sifive_x280_intr,
BLIS_COPYV_KER, BLIS_DCOMPLEX, bli_zcopyv_sifive_x280_intr,

BLIS_DOTV_KER, BLIS_FLOAT, bli_sdotv_sifive_x280_intr,
BLIS_DOTV_KER, BLIS_DOUBLE, bli_ddotv_sifive_x280_intr,
Expand All @@ -84,15 +84,15 @@ void bli_cntx_init_sifive_x280( cntx_t* cntx )
BLIS_DOTXV_KER, BLIS_SCOMPLEX, bli_cdotxv_sifive_x280_intr,
BLIS_DOTXV_KER, BLIS_DCOMPLEX, bli_zdotxv_sifive_x280_intr,

BLIS_INVERTV_KER, BLIS_FLOAT, bli_sinvertv_sifive_x280_asm,
BLIS_INVERTV_KER, BLIS_DOUBLE, bli_dinvertv_sifive_x280_asm,
BLIS_INVERTV_KER, BLIS_SCOMPLEX, bli_cinvertv_sifive_x280_asm,
BLIS_INVERTV_KER, BLIS_DCOMPLEX, bli_zinvertv_sifive_x280_asm,
BLIS_INVERTV_KER, BLIS_FLOAT, bli_sinvertv_sifive_x280_intr,
BLIS_INVERTV_KER, BLIS_DOUBLE, bli_dinvertv_sifive_x280_intr,
BLIS_INVERTV_KER, BLIS_SCOMPLEX, bli_cinvertv_sifive_x280_intr,
BLIS_INVERTV_KER, BLIS_DCOMPLEX, bli_zinvertv_sifive_x280_intr,

BLIS_INVSCALV_KER, BLIS_FLOAT, bli_sinvscalv_sifive_x280_asm,
BLIS_INVSCALV_KER, BLIS_DOUBLE, bli_dinvscalv_sifive_x280_asm,
BLIS_INVSCALV_KER, BLIS_SCOMPLEX, bli_cinvscalv_sifive_x280_asm,
BLIS_INVSCALV_KER, BLIS_DCOMPLEX, bli_zinvscalv_sifive_x280_asm,
BLIS_INVSCALV_KER, BLIS_FLOAT, bli_sinvscalv_sifive_x280_intr,
BLIS_INVSCALV_KER, BLIS_DOUBLE, bli_dinvscalv_sifive_x280_intr,
BLIS_INVSCALV_KER, BLIS_SCOMPLEX, bli_cinvscalv_sifive_x280_intr,
BLIS_INVSCALV_KER, BLIS_DCOMPLEX, bli_zinvscalv_sifive_x280_intr,

BLIS_SCAL2V_KER, BLIS_FLOAT, bli_sscal2v_sifive_x280_intr,
BLIS_SCAL2V_KER, BLIS_DOUBLE, bli_dscal2v_sifive_x280_intr,
Expand All @@ -104,20 +104,20 @@ void bli_cntx_init_sifive_x280( cntx_t* cntx )
BLIS_SCALV_KER, BLIS_SCOMPLEX, bli_cscalv_sifive_x280_intr,
BLIS_SCALV_KER, BLIS_DCOMPLEX, bli_zscalv_sifive_x280_intr,

BLIS_SETV_KER, BLIS_FLOAT, bli_ssetv_sifive_x280_asm,
BLIS_SETV_KER, BLIS_DOUBLE, bli_dsetv_sifive_x280_asm,
BLIS_SETV_KER, BLIS_SCOMPLEX, bli_csetv_sifive_x280_asm,
BLIS_SETV_KER, BLIS_DCOMPLEX, bli_zsetv_sifive_x280_asm,
BLIS_SETV_KER, BLIS_FLOAT, bli_ssetv_sifive_x280_intr,
BLIS_SETV_KER, BLIS_DOUBLE, bli_dsetv_sifive_x280_intr,
BLIS_SETV_KER, BLIS_SCOMPLEX, bli_csetv_sifive_x280_intr,
BLIS_SETV_KER, BLIS_DCOMPLEX, bli_zsetv_sifive_x280_intr,

BLIS_SUBV_KER, BLIS_FLOAT, bli_ssubv_sifive_x280_intr,
BLIS_SUBV_KER, BLIS_DOUBLE, bli_dsubv_sifive_x280_intr,
BLIS_SUBV_KER, BLIS_SCOMPLEX, bli_csubv_sifive_x280_intr,
BLIS_SUBV_KER, BLIS_DCOMPLEX, bli_zsubv_sifive_x280_intr,

BLIS_SWAPV_KER, BLIS_FLOAT, bli_sswapv_sifive_x280_asm,
BLIS_SWAPV_KER, BLIS_DOUBLE, bli_dswapv_sifive_x280_asm,
BLIS_SWAPV_KER, BLIS_SCOMPLEX, bli_cswapv_sifive_x280_asm,
BLIS_SWAPV_KER, BLIS_DCOMPLEX, bli_zswapv_sifive_x280_asm,
BLIS_SWAPV_KER, BLIS_FLOAT, bli_sswapv_sifive_x280_intr,
BLIS_SWAPV_KER, BLIS_DOUBLE, bli_dswapv_sifive_x280_intr,
BLIS_SWAPV_KER, BLIS_SCOMPLEX, bli_cswapv_sifive_x280_intr,
BLIS_SWAPV_KER, BLIS_DCOMPLEX, bli_zswapv_sifive_x280_intr,

BLIS_XPBYV_KER, BLIS_FLOAT, bli_sxpbyv_sifive_x280_intr,
BLIS_XPBYV_KER, BLIS_DOUBLE, bli_dxpbyv_sifive_x280_intr,
Expand All @@ -130,31 +130,31 @@ void bli_cntx_init_sifive_x280( cntx_t* cntx )
BLIS_AXPY2V_KER, BLIS_SCOMPLEX, bli_caxpy2v_sifive_x280_intr,
BLIS_AXPY2V_KER, BLIS_DCOMPLEX, bli_zaxpy2v_sifive_x280_intr,

BLIS_AXPYF_KER, BLIS_FLOAT, bli_saxpyf_sifive_x280_asm,
BLIS_AXPYF_KER, BLIS_DOUBLE, bli_daxpyf_sifive_x280_asm,
BLIS_AXPYF_KER, BLIS_SCOMPLEX, bli_caxpyf_sifive_x280_asm,
BLIS_AXPYF_KER, BLIS_DCOMPLEX, bli_zaxpyf_sifive_x280_asm,
BLIS_AXPYF_KER, BLIS_FLOAT, bli_saxpyf_sifive_x280_intr,
BLIS_AXPYF_KER, BLIS_DOUBLE, bli_daxpyf_sifive_x280_intr,
BLIS_AXPYF_KER, BLIS_SCOMPLEX, bli_caxpyf_sifive_x280_intr,
BLIS_AXPYF_KER, BLIS_DCOMPLEX, bli_zaxpyf_sifive_x280_intr,

BLIS_DOTXF_KER, BLIS_FLOAT, bli_sdotxf_sifive_x280_asm,
BLIS_DOTXF_KER, BLIS_DOUBLE, bli_ddotxf_sifive_x280_asm,
BLIS_DOTXF_KER, BLIS_SCOMPLEX, bli_cdotxf_sifive_x280_asm,
BLIS_DOTXF_KER, BLIS_DCOMPLEX, bli_zdotxf_sifive_x280_asm,
BLIS_DOTXF_KER, BLIS_FLOAT, bli_sdotxf_sifive_x280_intr,
BLIS_DOTXF_KER, BLIS_DOUBLE, bli_ddotxf_sifive_x280_intr,
BLIS_DOTXF_KER, BLIS_SCOMPLEX, bli_cdotxf_sifive_x280_intr,
BLIS_DOTXF_KER, BLIS_DCOMPLEX, bli_zdotxf_sifive_x280_intr,

BLIS_DOTAXPYV_KER, BLIS_FLOAT, bli_sdotaxpyv_sifive_x280_intr,
BLIS_DOTAXPYV_KER, BLIS_DOUBLE, bli_ddotaxpyv_sifive_x280_intr,
BLIS_DOTAXPYV_KER, BLIS_SCOMPLEX, bli_cdotaxpyv_sifive_x280_intr,
BLIS_DOTAXPYV_KER, BLIS_DCOMPLEX, bli_zdotaxpyv_sifive_x280_intr,

BLIS_DOTXAXPYF_KER, BLIS_FLOAT, bli_sdotxaxpyf_sifive_x280_asm,
BLIS_DOTXAXPYF_KER, BLIS_DOUBLE, bli_ddotxaxpyf_sifive_x280_asm,
BLIS_DOTXAXPYF_KER, BLIS_SCOMPLEX, bli_cdotxaxpyf_sifive_x280_asm,
BLIS_DOTXAXPYF_KER, BLIS_DCOMPLEX, bli_zdotxaxpyf_sifive_x280_asm,
BLIS_DOTXAXPYF_KER, BLIS_FLOAT, bli_sdotxaxpyf_sifive_x280_intr,
BLIS_DOTXAXPYF_KER, BLIS_DOUBLE, bli_ddotxaxpyf_sifive_x280_intr,
BLIS_DOTXAXPYF_KER, BLIS_SCOMPLEX, bli_cdotxaxpyf_sifive_x280_intr,
BLIS_DOTXAXPYF_KER, BLIS_DCOMPLEX, bli_zdotxaxpyf_sifive_x280_intr,

// Level 1m
BLIS_PACKM_KER, BLIS_FLOAT, bli_spackm_sifive_x280_asm_7m4,
BLIS_PACKM_KER, BLIS_DOUBLE, bli_dpackm_sifive_x280_asm_7m4,
BLIS_PACKM_KER, BLIS_SCOMPLEX, bli_cpackm_sifive_x280_asm_6m2,
BLIS_PACKM_KER, BLIS_DCOMPLEX, bli_zpackm_sifive_x280_asm_6m2,
BLIS_PACKM_KER, BLIS_FLOAT, bli_spackm_sifive_x280_intr,
BLIS_PACKM_KER, BLIS_DOUBLE, bli_dpackm_sifive_x280_intr,
BLIS_PACKM_KER, BLIS_SCOMPLEX, bli_cpackm_sifive_x280_intr,
BLIS_PACKM_KER, BLIS_DCOMPLEX, bli_zpackm_sifive_x280_intr,

// Level 3
BLIS_GEMM_UKR, BLIS_FLOAT, bli_sgemm_sifive_x280_asm_7m4,
Expand Down
2 changes: 1 addition & 1 deletion config/sifive_x280/make_defs.mk
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -61,7 +61,7 @@ endif
ifeq ($(DEBUG_TYPE),noopt)
COPTFLAGS := -O0
else
COPTFLAGS := -Ofast
COPTFLAGS := -O3
endif

# Flags specific to optimized kernels.
Expand Down
Loading