InstLatX64's Avatar

InstLatX64

@instlatx64

x86/x64, SIMD, #AVX512, "Aha!" moments. I have been writing code since 1986. Budapest, Europe https://instlatx64.github.io/InstLatx64/

278
Followers
0
Following
393
Posts
12.11.2024
Joined
Posts Following

Latest posts by InstLatX64 @instlatx64

#AMD refreshed the "AMD64 Architecture Programmer's Manual, Volumes 2" 24593 pdf to v3.44
docs.amd.com/v/u/en-US/24...

07.03.2026 09:27 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Post image Post image

#AMD has updated the "Seamless Firmware Maintenance (SFS) Specification" pdf 58604 to v0.76:
docs.amd.com/v/u/en-US/58...

07.03.2026 08:09 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

New #Intel #TDX docs:
Intel TDX Module Extension for Quoting
cdrdv2.intel.com/v1/dl/getCon...

04.03.2026 20:29 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

What CCD is used in #AMD #Sorano #EPYC8005? 84 cores is quite a strange number, 12*7?
www.amd.com/en/blogs/202...

03.03.2026 10:05 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Post image

#AMD released the "CPU Timing and Power Side Channel Guidance" pdf 71030 v1.1:
docs.amd.com/v/u/en-US/71...

01.03.2026 15:56 πŸ‘ 3 πŸ” 2 πŸ’¬ 0 πŸ“Œ 0

#AMD released the 7th #Zen6 pdf "Guest PMC Event Filtering" 69202 v1.00:
docs.amd.com/v/u/en-US/69...
CPUID Fn8000_000A_ECX[GuestPmcFiltering], bit 3
#Virtualization

28.02.2026 09:14 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Post image

#Intel SDE 10.7 Release Notes:
www.intel.com/content/www/...

27.02.2026 11:30 πŸ‘ 3 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

#Intel refreshed the #AVX10_2 specification to 6.0:
cdrdv2-public.intel.com/913918/36105...
#AVX10_VNNI_INT renamed to #AVX10_V1_AUX

27.02.2026 09:06 πŸ‘ 5 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

#Intel refreshed the Advanced Performance Extensions (#Intel #APX) Architecture Specification to 8.0:
cdrdv2-public.intel.com/913917/35582...

27.02.2026 08:11 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

#AMD released the 6th #Zen6 pdf "CPPC Performance Priority" 69206 v1.00:
docs.amd.com/v/u/en-US/69...
#CPPC

26.02.2026 11:45 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 1

#Intel released the SDE 10.7 (Software Development Emulator) with Indirect Branch History Fence ( #IBHF) instruction support and refreshed CPUID dumps:
www.intel.com/content/www/...
#NovaLake #DiamondRapids #APX #AVX10_2

26.02.2026 09:37 πŸ‘ 3 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0

#AMD released the 5th #Zen6 pdf "RMPOPT" 69203 v1.00:
docs.amd.com/v/u/en-US/69...
#RMPOPT #Virtualization

26.02.2026 09:01 πŸ‘ 3 πŸ” 0 πŸ’¬ 0 πŸ“Œ 1

#AMD released the 4th #Zen6 pdf "RMP Dirty" 69203 v1.00:
docs.amd.com/v/u/en-US/69...
#RMPCHKD #Virtualization

24.02.2026 09:08 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 1

#AMD has removed the public access to the #Zen6 "Performance Monitoring Counters for AMD Family 1Ah Model 50h-57h Processors" pdf.
Perhaps a new version reveals too much detail.
Bugzilla has archived v1.00:
bugzilla.kernel.org/attachment.c...

23.02.2026 08:57 πŸ‘ 3 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

#AMD has re-released PDFs 69191 and 69192, removing the "AMD Confidential" headers.
#AVX512_BMM specification:
docs.amd.com/v/u/en-US/69...
#FRED specification:
docs.amd.com/v/u/en-US/69...

22.02.2026 09:05 πŸ‘ 5 πŸ” 2 πŸ’¬ 0 πŸ“Œ 0
Post image Post image Post image Post image

This snow report was the inspiration for this #AMD #Zen codename table. Sources included.
#Zen2 #Zen3 #Zen4 #Zen5 #Zen6 #Zen7

19.02.2026 20:39 πŸ‘ 1 πŸ” 1 πŸ’¬ 0 πŸ“Œ 0

There are a few working #WildCatLake among #Intel test machines: (CPUID D06D1, 6c/6t (2P+4LPE probably), 2500 MHz base freq, no HTT, no AVX512, I think Intel 18A)
intel-gfx-ci.01.org/tree/intel-x...
#CougarCove #Darkmont

14.02.2026 09:22 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

According to this LLVM patch, #AMD CPUID BD0F00 is #Annapurna, a #Zen5 based #EPYC Embedded for Network Control Planes:
github.com/llvm/llvm-pr...

13.02.2026 21:34 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

#Intel #Xeon600 Processors for Workstation (GNR-W) - Turbo Tables:
www.intel.com/content/www/...
#GraniteRapidsW #GNRW

13.02.2026 09:18 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Post image

#Intel microcode refresh 20260211:
github.com/intel/Intel-...
CSV:
github.com/intel/Intel-...

11.02.2026 08:36 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

#Intel released the 90th edition of the Software Developer’s Manuals with canonized #FRED, #ArchitecturalPEBS, Asymmetric RDT-M & RDT-A
All-in-One:
cdrdv2-public.intel.com/874240/32546...
Changes v82:
cdrdv2-public.intel.com/874239/25204...
#LKGS, #PBNDKB

10.02.2026 07:21 πŸ‘ 3 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0
Preview
128-Core AMD EPYC 9754 (Bergamo, Zen4c) AA0F02 CPUID dump Β· InstLatx64/InstLatx64@35e0064 -- 128-Core AMD EPYC 9754 (Bergamo, Zen4c) AA0F02 CPUID dump added -- AMD Venice CPUID B50F00, B90F00 Dense info added -- AMD Venice CPUID BA0F00, BC0F00 Classic info added -- Intel Dunlow codename...

Github:
github.com/InstLatx64/I...
Sources:
[1]: x.com/x86deadandba...
[2]: www.youtube.com/watch?v=bO5g...
2/2

08.02.2026 09:28 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

New CPUID dump:
-- 128-Core #AMD #EPYC 9754 ( #Bergamo top-SKU, #Zen4c) AA0F02
New Intel codenames:
- #Dunlow (LGA1954 entry server [1])
- #TitanLake, #HammerLake [2]
- LGA1954 socket info #NovaLake, #RazorLake [2]
#AMD #Venice info:
- B50F00, B90F00 = Dense
- BA0F00, BC0F00 = Classic
1/2

08.02.2026 09:27 πŸ‘ 1 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0
Post image

Has anyone come across the pin number of the #AMD #SP8 socket? Maybe e.g. a cooler or motherboard manufacturer has already posted it somewhere...

04.02.2026 20:51 πŸ‘ 0 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

SDE 10.5.0 (Software Development Emulator) Release Notes:

03.02.2026 09:24 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

Fortunately, all 11 #Intel #GraniteRapidsW SKUs use dual FMA:
www.intel.com/content/www/...

03.02.2026 09:09 πŸ‘ 1 πŸ” 0 πŸ’¬ 0 πŸ“Œ 1
Post image Post image

Official #Intel #GraniteRapidsW SKU list:
www.intel.com/content/www/...
#GNR_W #AVX512 #AVX10_1 #AMX_FP16

02.02.2026 09:46 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 1

#Intel released the SDE 10.5.0 (Software Development Emulator), sync with 60th ISA Ext Guide:
www.intel.com/content/www/...
#NovaLake #DiamondRapids #APX #AVX10_2

01.02.2026 19:24 πŸ‘ 3 πŸ” 0 πŸ’¬ 0 πŸ“Œ 2

CC:
@geofflangdale.bsky.social

31.01.2026 20:06 πŸ‘ 2 πŸ” 0 πŸ’¬ 0 πŸ“Œ 0

One possible application of the #AVX512_BMM VPMACOR16x16x16 instruction is the compact index2msk function: if src2 is a full mask and src3 contains variably shifted masks, then VPMACOR16x16x16 will collect them into a single word. The order and repetitions are arbitrary.

31.01.2026 19:57 πŸ‘ 2 πŸ” 0 πŸ’¬ 1 πŸ“Œ 0