Gabriel
@dssgabriel.bsky.social
PhD candidate, HPC Software Engineering @cea.fr
MSc HPC & Simulation from @univparissaclay.bsky.social
Architecture, microbenchmarking & SIMD sorcery. Research on distributed computing, data structures & memory layouts at exascale.
RTFM 👹
MSc HPC & Simulation from @univparissaclay.bsky.social
Architecture, microbenchmarking & SIMD sorcery. Research on distributed computing, data structures & memory layouts at exascale.
RTFM 👹
Reposted by Gabriel
We’re delighted to welcome Modules to the High Performance Software Foundation as an established project 🎊
Read the announcement ➡️ hpsf.io/blog/2025/hi...
Read the announcement ➡️ hpsf.io/blog/2025/hi...
November 10, 2025 at 2:25 PM
We’re delighted to welcome Modules to the High Performance Software Foundation as an established project 🎊
Read the announcement ➡️ hpsf.io/blog/2025/hi...
Read the announcement ➡️ hpsf.io/blog/2025/hi...
Reposted by Gabriel
#AMD #Zen6 znver6 ISA:
- #AVX512_BMM (CPUID.80000021.EAX[23], VBMACOR16x16x16, VBMACXOR16x16x16, VBITREV)
- #AVX512_FP16
- #AVX_NE_CONVERT
- #AVX_IFMA
- #AVX_VNNI_INT8
Source:
sourceware.org/pipermail/bi...
- #AVX512_BMM (CPUID.80000021.EAX[23], VBMACOR16x16x16, VBMACXOR16x16x16, VBITREV)
- #AVX512_FP16
- #AVX_NE_CONVERT
- #AVX_IFMA
- #AVX_VNNI_INT8
Source:
sourceware.org/pipermail/bi...
November 8, 2025 at 1:11 AM
#AMD #Zen6 znver6 ISA:
- #AVX512_BMM (CPUID.80000021.EAX[23], VBMACOR16x16x16, VBMACXOR16x16x16, VBITREV)
- #AVX512_FP16
- #AVX_NE_CONVERT
- #AVX_IFMA
- #AVX_VNNI_INT8
Source:
sourceware.org/pipermail/bi...
- #AVX512_BMM (CPUID.80000021.EAX[23], VBMACOR16x16x16, VBMACXOR16x16x16, VBITREV)
- #AVX512_FP16
- #AVX_NE_CONVERT
- #AVX_IFMA
- #AVX_VNNI_INT8
Source:
sourceware.org/pipermail/bi...
Reposted by Gabriel
NASM 3.00 Assembler Is Ready With Intel APX & AVX10 Support - https://www.phoronix.com/news/NASM-3.00-APX-AVX10
NASM 3.00 Assembler Is Ready With Intel APX & AVX10 Support
Slipping under my radar in October was the release of NASM 3.00 and the follow-up NASM 3.01 release shortly there after. This widely-used open-source assembler is now ready with support for Intel's Advanced Performance Extensions (APX) and AVX10...
www.phoronix.com
November 2, 2025 at 1:53 PM
NASM 3.00 Assembler Is Ready With Intel APX & AVX10 Support - https://www.phoronix.com/news/NASM-3.00-APX-AVX10
Reposted by Gabriel
Reposted by Gabriel
#anniversaire 🎂 | Le CEA fête ses 80 🕯️ .
Le 18/10/1945, le Général de Gaulle signe l’ordonnance fondatrice du CEA.
⚛️L'organisme a pour ambition d’offrir à la 🇫🇷 la maîtrise de l’atome dans différents domaines de la science, de l’industrie & de la défense nationale
👇Extrait
Le 18/10/1945, le Général de Gaulle signe l’ordonnance fondatrice du CEA.
⚛️L'organisme a pour ambition d’offrir à la 🇫🇷 la maîtrise de l’atome dans différents domaines de la science, de l’industrie & de la défense nationale
👇Extrait
October 18, 2025 at 6:58 AM
#anniversaire 🎂 | Le CEA fête ses 80 🕯️ .
Le 18/10/1945, le Général de Gaulle signe l’ordonnance fondatrice du CEA.
⚛️L'organisme a pour ambition d’offrir à la 🇫🇷 la maîtrise de l’atome dans différents domaines de la science, de l’industrie & de la défense nationale
👇Extrait
Le 18/10/1945, le Général de Gaulle signe l’ordonnance fondatrice du CEA.
⚛️L'organisme a pour ambition d’offrir à la 🇫🇷 la maîtrise de l’atome dans différents domaines de la science, de l’industrie & de la défense nationale
👇Extrait
Reposted by Gabriel
Colleges do a terrible job of teaching C++.
It’s not “C with Classes”. Injected into curriculums as a demonstration of early CS concepts, it leaves many with a sour taste.
Students later immediately fall in love with the first language that *doesn’t* feel that way.
It’s not “C with Classes”. Injected into curriculums as a demonstration of early CS concepts, it leaves many with a sour taste.
Students later immediately fall in love with the first language that *doesn’t* feel that way.
October 13, 2025 at 9:21 PM
Colleges do a terrible job of teaching C++.
It’s not “C with Classes”. Injected into curriculums as a demonstration of early CS concepts, it leaves many with a sour taste.
Students later immediately fall in love with the first language that *doesn’t* feel that way.
It’s not “C with Classes”. Injected into curriculums as a demonstration of early CS concepts, it leaves many with a sour taste.
Students later immediately fall in love with the first language that *doesn’t* feel that way.
Reposted by Gabriel
HPSF Board Member & Kokkos Project Maintainer, Damien Lebrun-Grandié, of Oak Ridge National Laboratory (ORNL) will share a maintainer's perspective on Sustainable HPC Software in #HPC Best Practices Webinar - Oct 15 at 1:00pm EDT 🔎 Learn more:
ideas-productivity.org/events/hpcbp...
ideas-productivity.org/events/hpcbp...
October 7, 2025 at 7:54 PM
HPSF Board Member & Kokkos Project Maintainer, Damien Lebrun-Grandié, of Oak Ridge National Laboratory (ORNL) will share a maintainer's perspective on Sustainable HPC Software in #HPC Best Practices Webinar - Oct 15 at 1:00pm EDT 🔎 Learn more:
ideas-productivity.org/events/hpcbp...
ideas-productivity.org/events/hpcbp...
Reposted by Gabriel
Defer: Resource cleanup in C with GCCs magic
oshub.org/projects/ret...
oshub.org/projects/ret...
Defer: Resource cleanup in C with GCCs magic on OSHub
[Defer Macro]
“Warning: This is experimental, relies on GCC-specific extensions (__attribute__((cleanup)) and nested functions), and is not portabl...
oshub.org
October 1, 2025 at 1:14 PM
Defer: Resource cleanup in C with GCCs magic
oshub.org/projects/ret...
oshub.org/projects/ret...
Reposted by Gabriel
About a month ago, I gave a talk at the Rust Amsterdam meetup about Sguaba (the type-safe spatial math Rust crate), and the recording of that is now online for anyone who wants their head to hurt with frames of reference and coordinate transforms 😅
youtu.be/kESBAiTYMoQ
youtu.be/kESBAiTYMoQ
Sguaba: Type-safe spatial math in Rust
YouTube video by Jon Gjengset
youtu.be
September 29, 2025 at 2:26 PM
About a month ago, I gave a talk at the Rust Amsterdam meetup about Sguaba (the type-safe spatial math Rust crate), and the recording of that is now online for anyone who wants their head to hurt with frames of reference and coordinate transforms 😅
youtu.be/kESBAiTYMoQ
youtu.be/kESBAiTYMoQ
Reposted by Gabriel
Hello you fine Internet folks,
Today we are taking a look Intel's Xeon 6 Memory Subsystem and the changes that Intel made in order to fit up to 128 cores in a single CPU.
Hope y'all enjoy!
chipsandcheese.com/p/a-look-int...
old.chipsandcheese.com/2025/09/26/a...
Today we are taking a look Intel's Xeon 6 Memory Subsystem and the changes that Intel made in order to fit up to 128 cores in a single CPU.
Hope y'all enjoy!
chipsandcheese.com/p/a-look-int...
old.chipsandcheese.com/2025/09/26/a...
A Look into Intel Xeon 6’s Memory Subsystem
Intel’s server dominance has been shaken by high core count competition from the likes of AMD and Arm.
chipsandcheese.com
September 26, 2025 at 4:44 PM
Hello you fine Internet folks,
Today we are taking a look Intel's Xeon 6 Memory Subsystem and the changes that Intel made in order to fit up to 128 cores in a single CPU.
Hope y'all enjoy!
chipsandcheese.com/p/a-look-int...
old.chipsandcheese.com/2025/09/26/a...
Today we are taking a look Intel's Xeon 6 Memory Subsystem and the changes that Intel made in order to fit up to 128 cores in a single CPU.
Hope y'all enjoy!
chipsandcheese.com/p/a-look-int...
old.chipsandcheese.com/2025/09/26/a...
Reposted by Gabriel
This is a font joke.
September 20, 2025 at 8:19 PM
This is a font joke.
Reposted by Gabriel
Personal achievement: Relentlessly correcting Exaflops and the likes to ExaFLOP/s, and seeing that it slowly sticks, also in articles of the press 💪 #HPC
September 6, 2025 at 8:17 AM
Personal achievement: Relentlessly correcting Exaflops and the likes to ExaFLOP/s, and seeing that it slowly sticks, also in articles of the press 💪 #HPC
Reposted by Gabriel
Intel gives some more details on #ClearwaterForest at @hotchipsorg #hc2025.
- wider instruction decoders
- better branch prediction
- massive parallel OoO engine
- deeper execution engine
- optimized memory subsystem
hardwareluxx.de/index.php/ne...
- wider instruction decoders
- better branch prediction
- massive parallel OoO engine
- deeper execution engine
- optimized memory subsystem
hardwareluxx.de/index.php/ne...
August 25, 2025 at 4:05 PM
Intel gives some more details on #ClearwaterForest at @hotchipsorg #hc2025.
- wider instruction decoders
- better branch prediction
- massive parallel OoO engine
- deeper execution engine
- optimized memory subsystem
hardwareluxx.de/index.php/ne...
- wider instruction decoders
- better branch prediction
- massive parallel OoO engine
- deeper execution engine
- optimized memory subsystem
hardwareluxx.de/index.php/ne...
Reposted by Gabriel
Those who have been around #HPC a while know the historical significance (and probably the names) of Japan's flagship supercomputers. They always push the boundaries of technology, but in ways that are driven uniquely by the needs of Japan.
blogs.nvidia.com/blog/fugakun...
blogs.nvidia.com/blog/fugakun...
RIKEN, Japan’s Leading Science Institute, Taps Fujitsu and NVIDIA for Next Flagship Supercomputer
Japan is once again building a landmark high-performance computing system — not simply by chasing speed, but by rethinking how technology can best serve the nation’s most urgent scientific needs. At t...
blogs.nvidia.com
August 22, 2025 at 3:20 AM
Those who have been around #HPC a while know the historical significance (and probably the names) of Japan's flagship supercomputers. They always push the boundaries of technology, but in ways that are driven uniquely by the needs of Japan.
blogs.nvidia.com/blog/fugakun...
blogs.nvidia.com/blog/fugakun...
Reposted by Gabriel
Lots of details on FugakuNEXT were just announced (www.riken.jp/pr/news/2025...). NVIDIA GPUs, Fujitsu Monaka-X ARM CPUs. Big shift away from the path paved by Fugaku; I guess there's no point in doing anything but NVIDIA anymore!
#HPC
#HPC
理化学研究所、富士通およびNVIDIAとの国際連携による「富岳NEXT」開発体制を始動
理化学研究所（五神真理事長、以下「理研」）は、富士通株式会社（時田 隆仁 代表取締役社長、以下「富士通」）、NVIDIA Corporation（ジェンスン・ファンCEO、以下「NVIDIA」）とともに、スーパーコンピュータ「富岳」の次世代となる新たなフラッグシップシステム（開発コードネーム：「富岳NEXT」）に関して、理研を開発主体とした国際連携により設計および開発を開始することとしました。
www.riken.jp
August 22, 2025 at 1:29 AM
Lots of details on FugakuNEXT were just announced (www.riken.jp/pr/news/2025...). NVIDIA GPUs, Fujitsu Monaka-X ARM CPUs. Big shift away from the path paved by Fugaku; I guess there's no point in doing anything but NVIDIA anymore!
#HPC
#HPC
Reposted by Gabriel
If you're like me and can't remember if it's RCX or R8 for the fourth parameter... Get the Compiler Explorer ABI Mug: actual calling conventions for x86-64 System V, Windows, and ARM64. Support @godbolt.org: grab yours here -> shop.compiler-explorer.com/collections/...
August 17, 2025 at 10:02 PM
If you're like me and can't remember if it's RCX or R8 for the fourth parameter... Get the Compiler Explorer ABI Mug: actual calling conventions for x86-64 System V, Windows, and ARM64. Support @godbolt.org: grab yours here -> shop.compiler-explorer.com/collections/...
Reposted by Gabriel
Reposted by Gabriel
Going faster than GLIBC memcpy?
squadrick.dev/journal/goin...
squadrick.dev/journal/goin...
Going faster than memcpy
While profiling Shadesmar a couple of weeks ago, I noticed that for large binary unserialized messages (>512kB) most of the execution time is spent doing copying the message (using memcpy) between pro...
squadrick.dev
August 11, 2025 at 4:56 PM
Going faster than GLIBC memcpy?
squadrick.dev/journal/goin...
squadrick.dev/journal/goin...
Reposted by Gabriel
Reposted by Gabriel
I just finished my first meaningful #APX code path: #SHA3 / #Keccak1600 on 32 GPR registers: Only 130 uops + a few MOVs. I wonder if it can beat the latency of 35+ uop #AVX512 version on #Intel #DiamondRapids.
#Novalake , #PantherCove , #CoyoteCove
#Novalake , #PantherCove , #CoyoteCove
July 31, 2025 at 3:13 PM
I just finished my first meaningful #APX code path: #SHA3 / #Keccak1600 on 32 GPR registers: Only 130 uops + a few MOVs. I wonder if it can beat the latency of 35+ uop #AVX512 version on #Intel #DiamondRapids.
#Novalake , #PantherCove , #CoyoteCove
#Novalake , #PantherCove , #CoyoteCove
Reposted by Gabriel
"throw" expression has type "void", but has one special case if it is one of the cases in a ternary "?:" expression, where its void is "converted" to the other operand. This is the only situation where this happens, but it could be extended arbitrarily for return, break, continue, and goto.
July 26, 2025 at 6:31 PM
"throw" expression has type "void", but has one special case if it is one of the cases in a ternary "?:" expression, where its void is "converted" to the other operand. This is the only situation where this happens, but it could be extended arbitrarily for return, break, continue, and goto.
Reposted by Gabriel
July 26, 2025 at 4:59 PM