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

AMD Ryzen 9 3950X Review

anandtech.com

Post reply on HN
Print view View original

221–230 of 290 posts

Re: AMD Ryzen 9 3950X Review

#221

Post by mtarnovan »

lliamander wrote: post #205
mtarnovan wrote: post #199

I'm very curious how this CPU performs for Elixir compilation.

Indeed, the only compiler benchmark one typically sees are for C/C++. But Lots of developers use other languages, and it would be nice to have a cross-section because each language is so different. BTW, how much does elixir benefit from multiple cores for compilation?

I haven't done any rigorous testing, but looks compiling the project I'm working on does use all CPUs. There's also this: https://hexdocs.pm/elixir/Kernel.ParallelCompiler.html

Re: AMD Ryzen 9 3950X Review

#222

Post by IanCutress »

ibobev wrote: post #72

As a developer I am keen to see some compilation benchmarks. Unfortunately those kinds of benchmarks are almost never included in such reviews. Instead there are many gaming benchmarks which purpose is not exactly clear for me, after obviously gaming is not the primary target market for R9 3950X.

Normally I run a Chrome compile benchmark, but for whatever reason it wasn't running properly on Win 10 1909. When I get a chance to debug (55k miles of travel over the next four weeks), I'm going to see if I can fix it and expand that bit of our testing.

Re: AMD Ryzen 9 3950X Review

#223

Post by paulmd »

NicoJuicy wrote: post #218
paulmd wrote: post #211

Earlier quoted context omitted.

Do remember that Meltdown and Spectre affected literally everyone else in the industry (including POWER, ARM, and SPARC) and it's only by chance that AMD's branch predictor happens to be relatively difficult to train to follow a predictable path (which is a necessary condition for some of these exploits). Intel isn't cheating any more than anyone else, AMD got lucky. Cascade Lake and the latest Coffee Lake (and Comet…

So, AMD got lucky 218 times? https://windowsreport.com/amd-intel-security-vulnerabilities...

Most of those exploits are software/driver vulnerabilities, and AMD has plenty of their own. Like when Threadripper drivers installed an out-of-date web server running with SYSTEM privileges, or when a series of PSP and BIOS bugs let an attacker escalate from VM guest to control of the PSP and BIOS persistence. If you're not finding vulnerabilities in your software, you're not looking.

Intel is in a lot more verticals than AMD is - networking, storage, etc - and all of those get their own CVE number. But that's a more nuanced approach than "big number bad" isn't it?

Specifically, Intel is finding a lot because they're doing an enterprise-wide initiative to lock down their IPs. Is AMD making the same effort, or are there more vulnerabilities in those GPU drivers or in the chipset drivers that they're not looking for? After all, a year ago they were at the stage where they were installing an out-of-date Apache instance along with your motherboard drivers... doesn't sound like security is priority #1 for them.

Again, low number sounds good. But if you're not looking, it's actually not good.

And researchers are not looking on AMD much either. AMD is less than 5% of the server market right now, you get the big headlines from finding vulnerabilities in Intel now, not AMD. Give it another 5 years and there will be more scrutiny on AMD's hardware/software.

Re: AMD Ryzen 9 3950X Review

#224

Post by IanCutress »

mmastrac wrote: post #74
ibobev wrote: post #72

As a developer I am keen to see some compilation benchmarks. Unfortunately those kinds of benchmarks are almost never included in such reviews. Instead there are many gaming benchmarks which purpose is not exactly clear for me, after obviously gaming is not the primary target market for R9 3950X.

I believe that some of the YouTube channels have started including compilation benchmarks. GamersNexus [1] and Linus Tech Tips [2] both do! [1] https://www.gamersnexus.net/guides/3460-new-cpu-bench-method... [2] https://youtu.be/stM2CPF9YAY?t=4m49s

Normally we at AnandTech run a Chrome compile benchmark, but for whatever reason it wasn't running properly on Win 10 1909. When I get a chance to debug (55k miles of travel over the next four weeks), I'm going to see if I can fix it and expand that bit of our testing.

Re: AMD Ryzen 9 3950X Review

#225

Post by IanCutress »

lliamander wrote: post #202
ibobev wrote: post #72

As a developer I am keen to see some compilation benchmarks. Unfortunately those kinds of benchmarks are almost never included in such reviews. Instead there are many gaming benchmarks which purpose is not exactly clear for me, after obviously gaming is not the primary target market for R9 3950X.

As a developer, other benchmarks I would like to see: * compile benchmarks for different mainstream languages (Java, C#, etc.) * IDE related benchmarks (i.e. how long does it take to index a large solution/workspace) * source control conversion tests (svn to git). Obviously not something that happens everyday, but I've done it at two different jobs, and generally when it happens you are often converting many repos as…

Got any standardized methods that could be automated/scripted under a clean Windows environment and doesn't require internet access / licensing? Give me a shout - ian@anandtech.com

Re: AMD Ryzen 9 3950X Review

#226

Post by paulmd »

Retric wrote: post #220
paulmd wrote: post #211

Earlier quoted context omitted.

Do remember that Meltdown and Spectre affected literally everyone else in the industry (including POWER, ARM, and SPARC) and it's only by chance that AMD's branch predictor happens to be relatively difficult to train to follow a predictable path (which is a necessary condition for some of these exploits). Intel isn't cheating any more than anyone else, AMD got lucky. Cascade Lake and the latest Coffee Lake (and Comet…

Last I read AMD processors aren't affected by the Meltdown bug. And the overall performance hit was 9x as large for Intel. However, if that’s wrong I would love to read about it.

Like I said, AMD was unique on Meltdown in that their branch predictor uses a neural-net based approach that is difficult to train to follow a specific path. It affected everyone else in the industry - Intel, IBM, ARM, Oracle, everyone.

Re: AMD Ryzen 9 3950X Review

#227

Post by bob1029 »

neogodless wrote: post #12

This isn't the most groundbreaking release - it's not the first 16-core chip you can buy (edited), nor the first 7nm. No new clockspeed records were set. Still, $750 for all that power! Any professionals shopping for this, or waiting for 24 or 32-core Threadripper? Anyone trying to upgrade on an older motherboard, or are you getting a matching X570 to ensure maximum boost and PCI-e bandwidth?

I am waiting for the 3990X announcement before pulling the trigger on 3rd gen TR. I am already running a 2950X in my main workstation, and it's currently really hard to make an argument for even more performance... That said, I am targeting the 3970X as my next upgrade option, but AMD might be able to upsell me on the 3990X. Once the trigger has been pulled on 3rd gen, I will look at repurposing the 2950X workstation as a Jenkins build agent. We currently only have 2 vCPUs in AWS for Jenkins, so this could make a huge difference for our build process and give the old machine a really good 2nd life.

In terms of the 3990X specifically, I am most interested to see if they are going to provide additional platform capabilities. There were rumors regarding a TRX80/WRX80 platform, which seemed to imply to some tech journalists that there would be an octal channel variant of TR available early next year. It would be very hard to turn this down if it were an option.

Re: AMD Ryzen 9 3950X Review

#228

Post by IanCutress »

8K832d7tNmiQ wrote: post #153
kilo_bravo_3 wrote: post #64

Earlier quoted context omitted.

Are there any current products one can buy, regardless of expense, using Intel's 10nm? I know there was one product, the Core i3-8121U, a very middling performance chip which saw an extremely limited release. But it was discontinued.

That processor solely exists as PR stunt to satisfy investors. It has a faulty GPU, lower clock speed, higher temp, and should be considered as a lab sample instead. You can take a look at Charlie's article about it[1] [1]: https://semiaccurate.com/2018/05/29/is-intels-upcoming-10nm-...

AnandTech deep dive and review on the chip: https://www.anandtech.com/show/13405/intel-10nm-cannon-lake-...

Re: AMD Ryzen 9 3950X Review

#229

Post by IanCutress »

neogodless wrote: post #83
cracker_jacks wrote: post #56

Where is the 9980XE @ $979 coming from in the price vs performance chart? Where can I get a 9980XE for $979??

https://ark.intel.com/content/www/us/en/ark/products/189126/... Seems like it should be $1979 to match Intel's suggested pricing. (Retail looks like it's $1949 on sale, and typically even higher.)

Intel has announced the Cascade Lake-X i9-10980XE coming out later this month for $979. It's slightly higher in frequency,

Re: AMD Ryzen 9 3950X Review

#230

Post by Retric »

paulmd wrote: post #226
Retric wrote: post #220

Earlier quoted context omitted.

Last I read AMD processors aren't affected by the Meltdown bug. And the overall performance hit was 9x as large for Intel. However, if that’s wrong I would love to read about it.

Like I said, AMD was unique on Meltdown in that their branch predictor uses a neural-net based approach that is difficult to train to follow a specific path. It affected everyone else in the industry - Intel, IBM, ARM, Oracle, everyone.

If that’s what you meant then saying ‘Meltdown and Spectre affected literally everyone in the industry’ was literally false. Perhaps you could say almost everyone else in the industry in the future.

As to being lucky, I have no idea what their internal process is or what they considered. But luck seems like an odd way to talk about such things.

Post reply on HN