amd/blis
1
0
Fork 0
mirror of https://github.com/amd/blis.git synced 2026-05-11 09:39:59 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fixed out-of-bounds bug in sup s6x16m haswell kernel.

Browse Source 
Details:
- Fixed another out-of-bounds read access bug in the haswell sup
  assembly kernels. This bug is similar to the one fixed in 17b0caa
  and affects bli_sgemmsup_rv_haswell_asm_6x2m(). Thanks to Madeesh
  Kannan for reporting this bug (and a suitable fix) in #635.
- CREDITS file update.

Change-Id: I10ccf4d4f471d93e8c8cc4df422c686438fb04e9
This commit is contained in:
Field G. Van Zee
2022-07-26 17:29:32 -05:00
committed by Nallani Bhaskar
parent 4b1663213c
commit faff30b46a
2 changed files with 24 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
CREDITS
View File

@@ -46,6 +46,7 @@ but many others have contributed code and feedback, including
Matthew Honnibal @honnibal
Stefan Husmann @stefanhusmann
Francisco Igual @figual (Universidad Complutense de Madrid)
Madeesh Kannan @shadeMe
Tony Kelman @tkelman
Lee Killough @leekillough (Cray)
Mike Kistler @mkistler (IBM, Austin Research Laboratory)

41
kernels/haswell/3/sup/bli_gemmsup_rv_haswell_asm_s6x16m.c
View File

@@ -4475,34 +4475,39 @@ void bli_sgemmsup_rv_haswell_asm_6x2m
label(.SROWSTORED)
vfmadd231ps(mem(rcx, 0*32), xmm3, xmm4)
vmovsd(mem(rcx, 0*32), xmm0)
vfmadd231ps(xmm0, xmm3, xmm4)
vmovsd(xmm4, mem(rcx, 0*32))
add(rdi, rcx)
vfmadd231ps(mem(rcx, 0*32), xmm3, xmm6)
vmovsd(mem(rcx, 0*32), xmm0)
vfmadd231ps(xmm0, xmm3, xmm6)
vmovsd(xmm6, mem(rcx, 0*32))
add(rdi, rcx)
vfmadd231ps(mem(rcx, 0*32), xmm3, xmm8)
vmovsd(mem(rcx, 0*32), xmm0)
vfmadd231ps(xmm0, xmm3, xmm8)
vmovsd(xmm8, mem(rcx, 0*32))
add(rdi, rcx)
vfmadd231ps(mem(rcx, 0*32), xmm3, xmm10)
vmovsd(mem(rcx, 0*32), xmm0)
vfmadd231ps(xmm0, xmm3, xmm10)
vmovsd(xmm10, mem(rcx, 0*32))
add(rdi, rcx)
vfmadd231ps(mem(rcx, 0*32), xmm3, xmm12)
vmovsd(mem(rcx, 0*32), xmm0)
vfmadd231ps(xmm0, xmm3, xmm12)
vmovsd(xmm12, mem(rcx, 0*32))
add(rdi, rcx)
vfmadd231ps(mem(rcx, 0*32), xmm3, xmm14)
vmovsd(mem(rcx, 0*32), xmm0)
vfmadd231ps(xmm0, xmm3, xmm14)
vmovsd(xmm14, mem(rcx, 0*32))
//add(rdi, rcx)