Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Intel publishes misleading benchmarks against AMD

servethehome.com

Post reply on HN
Print view View original

201–209 of 209 posts

Re: Intel publishes misleading benchmarks against AMD

#201

Post by ksec »

bob1029 wrote: post #184
bjornjaja wrote: post #172

Earlier quoted context omitted.

AMD bringing the competition forces Intel to up their game and bolster R&D—they may come with some beast chips as a result. Such is the way things work, scales tip one way then the next. What about NVIDIA?? Also isn’t Amazon working on their own chips?

x86 licensing makes this an AMD vs Intel game, with occasional celebrity appearances by VIA, et. al. So, effectively AMD gets this private 1v1 game with Intel, and has a strong part in the GPU side of the market as well. GPU is certainly a more open playing field compared to x86 (Intel is trying to get into it), but AMD is very competitive here as well. I think we will start seeing a lot more unification of their con…

Hold on a min,

>Dr. Su said AMD was more about architecture than process technology, so perhaps they can prove what most of us suspected all along - Intellectual property is the most valuable strategic asset we have. Even if Intel reverse-engineered AMD's reticles, by the time they could implement the same architecture, AMD would be pushing products to market that are 2-3 generations ahead. My question right now is, how long will it take Intel to reach parity on this architecture front (especially if it's the key to success as AMD's CEO claims)?

This is not true. As much as I support AMD and loathe Intel, architecture wise I would say even Zen 2 and Intel SomethingLake is about the same, with many benchmarks backing the claim. And Icelake being slightly in front of Zen 2. Icelake did deliver a roughly 15% improvement as promised, so there is no sign Intel's architecture department has any problems ( yet ).

The problem is Icelake performance improvement doesn't matter if 10nm cant yield.

And when Dr Su said about IP, she was referring to the whole chipset strategy and how it blinds everything together.

Re: Intel publishes misleading benchmarks against AMD

#202

Post by zymhan »

NicoJuicy wrote: post #168
zymhan wrote: post #159

Earlier quoted context omitted.

Agreed, while AMD has a lot of market potential, their stock is overvalued given current performance. I'm rooting for them, and I want to put money on them, but I sold my shares I got for <$10 and I can't imagine getting back in at over $30.

I was in the same situation ( 3 to 10 and then 25 to 36 now). It ain't going to stop untill 2022. And if you believe Intel, it's 2021 instead of 2022 ;) PS. Facts related: AMD Gains Chip Market Share in Nearly Every Category.

Sure, but it's really easy for them to post massive growth in the server category since they started with market share in the low single digits.

They have nowhere to go but up, but that can still mean their stock price is too high.

Re: Intel publishes misleading benchmarks against AMD

#203

Post by Smithalicious »

NicoJuicy wrote: post #39

They wouldn't have an answer to AMD untill 2021: https://newsroom.intel.com/news/2019-intel-investor-meeting/... Ps. Long AMD, perfect execution the last two years. It's crazy, even their late firmware fixes improve performance as it had bad press for a moment for false benchmarks advertisement ( it was not false so it seems, the firmware was just not ok yet. Not everything is fully ok now, it's the only minor nitpic…

As soon as people's posting style moves from "don't take investment advice from comments on the internet" to "long AMD" is when I start getting suspicious that I maybe shouldn't long AMD.

Granted I'm a penniless student so I'm not longing or shorting anything but you get the point

Re: Intel publishes misleading benchmarks against AMD

#204

Post by rhizome »

axaxs wrote: post #131
rhizome wrote: post #119

Earlier quoted context omitted.

You can get a VESA mount Mini-ITX[1] case with a current-chipset (X570, etc.) motherboard, as well as all the other Mini-ITX options like monitor-mount base cases[2]. 1. https://www.amazon.com/Antec-ISK110-VESA-U3-Mini-ITX-Compati... 2. https://encrypted-tbn0.gstatic.com/images?q=tbn%3AANd9GcQQB9...

Thanks for the links. This is significantly larger than a NUC, but still manageable. That said, part of what I appreciate about the NUC is that it's all built together...no weird hardware/compat issues, etc.

The weird hw/compat issues tend to be OS-related in my experience. I checked a few of these boards and the integrated graphics tended to be Radeon, but I think it'd really only matter under Linux.

Re: Intel publishes misleading benchmarks against AMD

#205

Post by sliken »

tagrun wrote: post #182
kllrnohj wrote: post #100

Earlier quoted context omitted.

> So, we're talking about desktops not used to browse the internet, and bare metal servers whose network stacks aren't exposed to hostile actors? To be fair the context here is HPC. Nobody is installing Chrome on something like IBM's Summit supercomputer. Nor is it running "untrusted" code. It's configured to go as fast as possible, nothing more. Anyone in cloud hosting or desktop would be insane to not install spect…

Except HPC clusters do need all the Spectre/Meltdown fixes and more. Those are shared environments (with the exception of a handful of people who got research grants for their own internal clusters). In fact, hyper-threading has also been disabled completely is some clusters.

I'm sure there are some, I'd be very surprised it it was most though.

The main problem with Spectre/Meltdown and friends is two VPSs under the same hypervisor sharing a core running some webstack. With careful analysis of tlb, cache timings, and related you can extract SSL private keys, bitcoin addresses, and similar information that would normally be quite hidden since you are running in a different kernels sharing the same hardware.

With a HPC cluster you are (generally) running on bare metal (no hypervisor), single linux kernel, and can see what the other user (if there is one) with simple tools like ps, w, top, and friends. Additionally there's generally not private keys for SSL or bitcoin addresses.

In fact often unencrypted network filesystems are used sending plain text over the wire. Even on more sensitive clusters the security is highest between the internet and the head node and less so between compute nodes.

Hyperthreading being on vs off is usually just an issue with the performance characteristics of whatever application is most common and the limitations of certain batch queues like sun grid engine (SGE). If your popular app hates hyperthreading then you turn it off. Or sometimes you want to minimize the performance impact of users sharing a core.

Additionally sometimes cores, or even nodes are not shared. But again it's for performance reasons not security. Last thing you want is a 10,000 core job to run 50% as fast because one node is shared with a resource intensive application.

Re: Intel publishes misleading benchmarks against AMD

#206

Post by sliken »

GhettoMaestro wrote: post #142
stagger87 wrote: post #124

Earlier quoted context omitted.

HPC is the use case in this thread, considering the topic is revolving around HPC benchmarks.

Let me clarify: When it comes to Spectre/Meltdown, HPC users and environments are most likely not prioritized targets of interest. The low hanging fruit in that scenario is stealing cryptographic keys from co-tenants on a public cloud.

Indeed, and generally HPC doesn't use hypervisors anyways.

Re: Intel publishes misleading benchmarks against AMD

#207

Post by sliken »

gpm wrote: post #67
sp332 wrote: post #60

Earlier quoted context omitted.

I'm not clear what your point was about the VMs. Why do hypervisors need Meltdown mitigated? I think you're underestimating the % of HPC networks with properly-configured firewalls and only running code from trusted sources.

If you trust all the code (and other inputs, or outputs) on all the VMs, and you are running it in a secure environment, you don't. I admit, it's possible I'm underestimating how common this is. Why you generally do need it is the above isn't true, and microcode has to be loaded at the hypervisor level for everyone or for no one

HPC doesn't generally use VMs or Hypervisors.

Re: Intel publishes misleading benchmarks against AMD

#208

Post by sliken »

robocat wrote: post #118
computerex wrote: post #48

Earlier quoted context omitted.

You absolutely need Spectre/Meltdown mitigation otherwise someone could attack you with JavaScript running in your web browser. https://linuxreviews.org/HOWTO_make_Linux_run_blazing_fast_(... noibrs - We don't need no restricted indirect branch speculation noibpb - We don't need no indirect branch prediction barrier either nospectre_v1 and nospectre_v2: Don't care if some program can get data from some other program…

> You absolutely need Spectre/Meltdown mitigation The article context is HPC - it seems very unlikely Spectre/Meltdown fixes are needed for HPC loads (no browser, no VM sharing).

Not only no VM sharing, but no VMs.

Re: Intel publishes misleading benchmarks against AMD

#209

Post by paulmd »

greatpatton wrote: post #144
paulmd wrote: post #45

Earlier quoted context omitted.

Also, the difference in sub-NUMA configuration actually favors AMD as well. NPS=4 is optimal for AMD, as shown by ServeTheHome's own work earlier. Both systems are tested in their proper (fastest) configuration. Despite ServeTheHome's own previous work showing this, they're whining about it being different. But if Intel hadn't tested AMD in the proper configuration they would have complained about that too. Yeah, you…

oh really one thread per core is the optimum for EPYC server? I don't know why AMD care about hyper threading then.

BTW turns out that was a typo and they used two threads per core anyway.
Post reply on HN