#zen4

Benjamin Carr, Ph.D. 👨🏻‍💻🧬BenjaminHCCarr@hachyderm.io
2025-02-26

Sizing up #MI300A’s #GPU
It’s well ahead of #Nvidia’s #H100 PCIe for just about every major category of 32- or 64-bit operations. MI300A can achieve 113.2 TFLOPS of #FP32 throughput, with each FMA counting as two floating point operations. For comparison, H100 PCIe achieved 49.3 TFLOPS in same test.
#AMD cut down #MI300X’s GPU to create MI300A. 24 #Zen4 cores is a lot of #CPU power, and occupies one quadrant on the MI300 chip. But MI300’s main attraction is still the GPU.
chipsandcheese.com/p/sizing-up

In Mediainmedia
2025-02-26

🔥 AMD Ryzen 9 7950X3D – Gaming & Création sans limites 🎮🚀

✅ 16 cœurs / 32 threads | 5.7 GHz Boost
✅ Zen 4 (5 nm) | 144 Mo AMD 3D V-Cache
✅ Multitâche & performances extrêmes

⚡ Jouez, créez et streamez avec fluidité !
🛒 Commandez ici: inmedia.ma/product/amd-ryzen-9

Benjamin Carr, Ph.D. 👨🏻‍💻🧬BenjaminHCCarr@hachyderm.io
2025-02-05

How to make any #AMD Zen #CPU always generate 4 from RDRAND
"We have demonstrated the ability to craft arbitrary malicious microcode patches on Zen 1 through #Zen4 CPUs," the Google #Security Team said.
"The #vulnerability is that the CPU uses an insecure hash function in the signature validation for #microcode updates."
The flaw listed as CVE-2024-56161 with a CVSS score of 7.2 out of 10, was discovered and reported to AMD in September, and a fix was devised by December.
theregister.com/2025/02/04/goo

Jeroen Ruigrok van der Wervenasmodai
2024-12-02
Benjamin Carr, Ph.D. 👨🏻‍💻🧬BenjaminHCCarr@hachyderm.io
2024-11-21

#AMD crafts custom #EPYC #CPU with #HBM3 memory for #Microsoft #Azure – CPU with 88 #Zen4 cores and 450GB of HBM3 may be repurposed #MI300C, four chips hit 7TB/s
#HBv series of Azure #VM are focused on delivering high amounts of memory bandwidth; Microsoft calls it “biggest #HPC bottleneck.” Previously, Microsoft had used #MilanX and #GenoaX with AMD 3D #VCache to provide extra bandwidth, but for latest #HBv5 VMs, Microsoft clearly wanted something even more performant.
tomshardware.com/pc-components

2024-11-16

In other words, Intel didn't design their architecture with in mind, so consumers will have to wait years for something similar.

What's worse, AMD has 3D V-Cache processors for $200, $300, $400 and $500. Regardless of benchmarks, X3D = gaming and Intel has no direct response.

videocardz.com/newz/intel-has-

2024-11-13

You know? I'm also targeting previous gen sales.

It's either a giant leap with , or just almost-same performance with the prior and laptops. I can trade efficiency if that's the case.

Multi Purr Puss :verified:platymew@layer8.space
2024-11-11

Oof, keeping the most recent #AMD platform driver performance kerfluffle on #Windows in mind, i might not be getting the best perf on my gaming-only dual-boot OS.

I've transplanted it from my long-gone/sold-off 5800X(#Zen3) to a #Zen4 (kinda-mobile) platform - 7840HS, with Radeon 780M (RDNA3 x12).

I'm daily-driving #Ubuntu 24.04 LTS, for non-gaming related workloads - i love this tiny box!

But what if i could get a few more FPS - FSR makes CP2077 a blocky mess, anyway - (re)install Win11?

2024-11-08

Honestly, I don't see the point of having a third-party going for more than $600. Makes no freaking sense.

After that price these compete in the space, where you can find superior SoC on all sides: , , or . These win hands down no contest.

Multi Purr Puss :verified:platymew@layer8.space
2024-10-02

I'd really like 16 CUs of #RDNA 3.5, instead of 12x RDNA3 in my 7840HS #AMD APU.

Meh, i chose my particular #miniPC because it's got a fancy #OCuLink port - an octa-core #Zen4 CPU is perfectly fine. My upgrade path has always been following the #eGPU route to 1080p60 max settings.

It's probably gonna be a #Radeon 7800, because of the preferable memory, capacity as well as bandwidth.

It's a somewhat annoying time of the year, October, with next-gen announcements at #CES looming. 😅

Benjamin Carr, Ph.D. 👨🏻‍💻🧬BenjaminHCCarr@hachyderm.io
2024-09-10

#Zen5 Variants and More, Clock for Clock
They compare #Zen 5 variants along with a pile of older cores with clock speed taken out of the equation as much as possible. That means capping clock speeds on all CPUs to the lowest clocking chip. In this case, that’s the #3GHz Athlon II X2 651.
Desktop Zen 5 takes a healthy 20.8% lead over #Zen4 in the most comparable configuration. #VCache gives Zen 4 96 MB of L3 cache which shrinks Zen 5’s lead to a still significant 16.6%.
chipsandcheese.com/2024/08/20/

WinFuture.deWinFuture
2024-09-05

Unerwartet ist der 5 ab sofort auch in erhältlich. Der - von mit 3D V-Cache wird exklusiv über nur einen Händler vertrieben. Die -Konkurrenz ist aber groß. winfuture.de/news,145049.html?

WinFuture.deWinFuture
2024-08-27

Die neuen - von enttäuschten zuletzt in . Mit einem - wird die -Leistung jetzt allerdings deutlich verbessert. Auch - profitieren. winfuture.de/news,144803.html?

ComputerBaseComputerBase
2024-08-20
WinFuture.deWinFuture
2024-07-23

Ende des Monats ist die neue -Generation von erhältlich. In einem - schlägt eine alte - jetzt allerdings den neuen der -Serie. winfuture.de/news,144096.html?

Benjamin Carr, Ph.D. 👨🏻‍💻🧬BenjaminHCCarr@hachyderm.io
2024-07-02

#AMD #Zen4 vs. #Intel Core Ultra 7 "#MeteorLake" In 400+ Benchmarks On Linux 6.10
The #CPU performance for the #Ryzen7 #7840U was around 27% faster than the Core Ultra 7 #155H while the #Ryzen 7 #7840HS was about 35% faster than that Meteor Lake laptop. AMD Zen 4 wins were all the more compelling when looking at power consumption with 7840U consuming 27.1W on average across all benchmarks with a peak of 51W compared to Core Ultra 7 having a 29.6W average and peak of 65 W.
phoronix.com/review/linux-610-

Benjamin Carr, Ph.D. 👨🏻‍💻🧬BenjaminHCCarr@hachyderm.io
2024-06-03

#AMD Previews “#Turin#Epyc #CPU, Expands Instinct GPU Roadmap
The 5th Gen AMD EPYC processors are built atop #Zen5 and will come in designs up to 192 cores / 384 threads per socket! Presumably the 192 core processors will be leveraging #Zen5C cores.
The top-end #Ryzen9000 has 16 cores and 32 threads and runs at 5.67 GHz turboed up, and the Zen 5 cores are averaging 16% more instructions per core than #Zen4.
nextplatform.com/2024/06/03/am

Benjamin Carr, Ph.D. 👨🏻‍💻🧬BenjaminHCCarr@hachyderm.io
2024-05-08

#AMD #Instinct #MI300A #APU With #CDNA3 #GPU, #Zen4 #CPU & #UnifiedMemory Offers Up To 4x Speedup Versus Discrete GPUs In #HPC
Since the AMD Instinct MI300A accelerator uses a unified #HBM interface, it eliminates the need for data replication and does not require a #programming distinction between the host and the device memory spaces.
wccftech.com/amd-instinct-mi30

Client Info

Server: https://mastodon.social
Version: 2025.04
Repository: https://github.com/cyevgeniy/lmst