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

231–240 of 290 posts

Re: AMD Ryzen 9 3950X Review

#231

Post by linkgoron »

seminatl wrote: post #208

I don't really get how they reached their conclusion. It seems like on most of the tests this new part gets beat by a cheaper one from Intel. It seems like a kinda unfair approach to use handbrake without AVX-512 support? Also not sure why they include the 3-d particle thing without AVX ... I guess because Intel is just too fast on that? If you look through the results, the things most people want to do with a comput…

You could just buy a CPU for a third or fourth of the price if you want to just browse. You don't need a $750 16-core CPU. You can buy a $200 9600KF or a $240 3600x or something even cheaper. For anything multi-core the 3950X blows the 9900k out of the water, and for 1440p gaming the 3950x trades blows with the 9900k. If you want 1080p gaming, or web-browsing, this is not the CPU for you. That said, even for 1080p gaming and single-threaded benchmarks the 3950 is usually within a few percent of the 9900K.

Re: AMD Ryzen 9 3950X Review

#232

Post by IanCutress »

bransonf wrote: post #73
davidy123 wrote: post #44

Slight meta, and part of a larger 'rant,' but when are we going to get away from reviews that might as well have been printed on glossy pages in PC Magazine in 1986? Anandtech has been at it since the 90s and hasn't changed their format at all. This is a serious decades-long stagnation of the web. The graphs should be dynamic (able to choose what scenarios and components to compare, and able to search within them), a…

I mean, there are plenty of options to compare new cpus to old ones. 0. CPUBoss 1. CPUBenchmark 2. UserBenchmark And if you’re considering an upgrade, it’s almost always the case that you should be comparing latest generation offerings to arrive at a purchase decision. Don’t call the web stagnant because they decide not to flood the page with JS libraries.

www.anandtech.com/Bench for real world benchmark data

Re: AMD Ryzen 9 3950X Review

#233

Post by snvzz »

rarecoil wrote: post #194
velox_io wrote: post #127

Earlier quoted context omitted.

Wow, I didn't expect that. Laptops were the only market Intel was still strong, still dominated. Having brought an rx5700 XT recently, the power usage has been just as impressive as the performance. AMD will have some very nice products when they start putting mini RX5000s in their CPUs. Decent performance without discrete GPUs is a real possibility. (Just need add some HBM (High Bandwidth Memory) to the mix and thin…

I bought an RX 5700 XT almost immediately after launch, and I can't say that I've had had much luck with RDNA/Navi. I experienced freezes/crashes in Windows 10 while gaming, and the card's hot, loud, and still unsupported by ROCm for compute on Linux. There are rumours that ROCm won't ever support this generation of Navi and is waiting on Arcturus[1]. The drivers still aren't there and support is limited outside of s…

>RX5700XT almost immediately after launch

So a reference card, which came with an horrendous, insufficient cooling solution.

The experience has been much better for those that bought the custom cards (such as sapphire pulse/nitro) later. They also got far more mature drivers to start with.

Re: AMD Ryzen 9 3950X Review

#234

Post by paulmd »

Damogran6 wrote: post #175
velox_io wrote: post #99

Earlier quoted context omitted.

The problem with this strategy is that they often have much lower clock speeds/ IPC, plus they are often quite power-hungry and lose the ability to overclock (that one's subjective). So you do often get bitten by Amdahl's law. However, if it is purely a rendering machine and single-threaded performance isn't an issue, you can get a lot of processing power for the money.

And potentially REALLY NOISY...

GamersNexus had a fun video playing with a new Epyc server with Wendell from L1Techs... the delta fans on that pull up to 7A. Not 0.7A, 7A. More wattage than your CPU.

https://youtu.be/la0_2Kmrr1E?t=164

Not an epyc thing, servers are hilariously loud and I've seen so many posts on selfhosted and homelab about "the wife made me get rid of my R710 because you can hear it in the closet/basement/etc all the time"....

Re: AMD Ryzen 9 3950X Review

#235

Post by seminatl »

linkgoron wrote: post #231
seminatl wrote: post #208

I don't really get how they reached their conclusion. It seems like on most of the tests this new part gets beat by a cheaper one from Intel. It seems like a kinda unfair approach to use handbrake without AVX-512 support? Also not sure why they include the 3-d particle thing without AVX ... I guess because Intel is just too fast on that? If you look through the results, the things most people want to do with a comput…

You could just buy a CPU for a third or fourth of the price if you want to just browse. You don't need a $750 16-core CPU. You can buy a $200 9600KF or a $240 3600x or something even cheaper. For anything multi-core the 3950X blows the 9900k out of the water, and for 1440p gaming the 3950x trades blows with the 9900k. If you want 1080p gaming, or web-browsing, this is not the CPU for you. That said, even for 1080p ga…

Ok but if it’s “trading blows with” and “within a few percent of” a CPU costing half the money, why is it rational to prefer it?

I know why I want AMD parts and that is because of ECC memory support but this article doesn’t even mention it.

Re: AMD Ryzen 9 3950X Review

#236

Post by seminatl »

gameswithgo wrote: post #215
seminatl wrote: post #208

I don't really get how they reached their conclusion. It seems like on most of the tests this new part gets beat by a cheaper one from Intel. It seems like a kinda unfair approach to use handbrake without AVX-512 support? Also not sure why they include the 3-d particle thing without AVX ... I guess because Intel is just too fast on that? If you look through the results, the things most people want to do with a comput…

avx2 performance between the new amd cpus and intel is comparable. ive done some detailed comparisons with my own code that is tuned for intel and the ryzen is quicker in some places, slower in others.

AVX-512 is not AVX2.

Re: AMD Ryzen 9 3950X Review

#237

Post by mtarnovan »

Bayart wrote: post #204
mtarnovan wrote: post #199

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

Compilation should profit pretty linearly from more cores thrown at it. Considering the Erlang VM is very concurrency-conscious it should be a pretty natural fit. But I'm not at all in on the intricacies of it, are there any particular CPU features (instruction set or module) it's known to take advantage of ?

You mean like vectorized instructions and such? I don't know, but if I had to guess I'd say no.

Anyway, I'm sure there are some compilation phases that won't benefit from more cores, but the bulk of the compilation seems to do so. Even though Elixir has incremental compilation, due to use of metaprogramming in the Phoenix framework sometimes even changes to a single file will trigger recompilation of hundreds of others, so I was thinking these new AMD CPUs with lots of cores could be very helpful here.

Re: AMD Ryzen 9 3950X Review

#238

Post by ajuc »

Bootwizard wrote: post #174
ajuc wrote: post #107

Earlier quoted context omitted.

US is far from average needs of a car user. Not a representative sample.

A country the size of all of Europe is not representative of the vehicle market? Not sure I buy that.

USA is very sparsely populated (35 people per square km, 174th in the world), most of the people live in countries with more than 100 people per square km.

To be exact 67% of people live in such countries, and 83% of people live in countries more dense than US.

Source: http://worldpopulationreview.com/countries/countries-by-dens... https://docs.google.com/spreadsheets/d/1aoJUhDHhhIvLRSPLsT7b...

Re: AMD Ryzen 9 3950X Review

#239

Post by tomp »

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.

Isn't the reason that AMD isn't affected, that they don't prefetch / cache pages before checking instruction's access privileges?

Re: AMD Ryzen 9 3950X Review

#240

Post by Matthias247 »

snvzz wrote: post #233
rarecoil wrote: post #194

Earlier quoted context omitted.

I bought an RX 5700 XT almost immediately after launch, and I can't say that I've had had much luck with RDNA/Navi. I experienced freezes/crashes in Windows 10 while gaming, and the card's hot, loud, and still unsupported by ROCm for compute on Linux. There are rumours that ROCm won't ever support this generation of Navi and is waiting on Arcturus[1]. The drivers still aren't there and support is limited outside of s…

>RX5700XT almost immediately after launch So a reference card, which came with an horrendous, insufficient cooling solution. The experience has been much better for those that bought the custom cards (such as sapphire pulse/nitro) later. They also got far more mature drivers to start with.

I would agree on bad cooling on reference cards. But regarding drivers and general I would not expect a lot of difference between all those cards.
Post reply on HN