Elektrine
Log in Register

#hpc

13 posts · Last used Sep 08

BuySellRam.com @BuySellRam@mstdn.business
· 2mo ago
AMD raised its 2030 data center CPU forecast to roughly $220 billion, then stated its EPYC performance claims in "agents per rack." The footnote admits agent counts are derived from available thread resources used as a proxy. Nvidia's Vera makes the opposite bet: 88 faster cores rather than 256 denser ones. Neither has published the benchmark that would decide which is right. https://www.buysellram.com/blog/agentic-server-cpu-wide-narrow-bets/ #DataCenter #ServerCPU #AgenticAI #AIInfrastructure #AMD #EPYC #Nvidia #Arm #Intel #HPC #Tech
buysellram.com
0
0
0
0
Dr. Moritz Lehmann @ProjectPhysX@mast.hpc.social
· 2mo ago
Replying to
@azonenberg@ioc.exchange @Methylzero@mast.hpc.social @niconiconi@mk.absturztau.be for #HPC workloads the low arithmetic intensity is just perfect. High memory bandwidth (1.5 TB/s) and capacity are the star of the show. Even with half the die size and compute, you would still get the same performance in lattice Boltzmann for example, so memory bound these codes are. Similar story for some AI inference workloads, and that's what drives the market price of pretty much all GPUs nowadays.
0
2
0
0
niconiconi @niconiconi@mk.absturztau.be
· 2mo ago
Replying to
It's official, the CMP 170HX performance has been unlocked to the Nvidia A100 level by a firmware exploit! 🚀 The community research basically started since three years ago after I publicly tested and reported this card in my blog post. I couldn't hack this card, but "a single spark can start a prairie fire." https://zenodo.org/records/20916112 PoC (mirror): https://github.com/fulracoco/cmpunlocker #hpc #gpu
zenodo.org
0
2
0
0
Gábor SAMU @gabor_samu@mstdn.ca
· 3mo ago
Using LSF Simulator with Summit (retired) job data, the study shows #AI workloads exhibit distinct scaling/runtime behavior, motivating scheduler and policy changes for efficient co-scheduling with large #MPI jobs. https://www.osti.gov/biblio/2205452 #HPC #IBM #LSF
osti.gov

Utilizing IBM Spectrum LSF Simulator to Understand the Impacts of Adding AI Workloads to Capability Supercomputing (Technical Report) | OSTI.GOV

Machine Learning and Artificial Intelligence has been identified as an emerging priority science area within the Department of Energy. Large scale accelerator based supercomputers like Summit, while traditionally employed for modeling and simulation, provide architectures that are suitable for accelerating the ML/AI workloads at scale. With the release of Summit in 2018, there was an increase in the number of ML/AI based projects seeking time on the machine. It quickly became apparent that the a

0
0
0
0
Self-Hosted Feed @selfhosted_bot@fd.mrmave.work
· 3mo ago
📚 trevor-vincent/awesome-high-performance-computing A curated list of awesome high performance computing resources Archives essential software libraries, hardware trends, and historical documentation for parallel computing system development ⭐ Stars: 1268 📅 Last Update: Jun 26, 2026 https://github.com/trevor-vincent/awesome-high-performance-computing #selfhosted #homelab #selfhost #selfhosting #opensource #hpc #library
github.com
0
0
0
0
kkarhan @kkarhan@jorts.horse
· 5mo ago
Replying to
@Infoseepage@mastodon.social @ChristosArgyrop@mastodon.social @thatdnaguy@genomic.social at least unlike #Shitcoin-#ASIC|s for the like of #Bitcoin these #GPU|s can at least be reused for something productive. Worst-case it'll make #Gaming, #CAD and #HPC "affordable" again.Best case we'll see actual #innovation outside of #Deepfake|s & #Disinformation campaigns!
0
0
0
0
You've seen all posts