Official AMD Ryzen Benchmarks, Reviews, Prices, and Discussion

[R0H1T]

Some amazing power charts ~ https://www.legitreviews.com/amd-ryzen-7-1800x-1700x-and-1700-processor-review_191753/12

 

[AtenRa]

Some amazing power charts ~ https://www.legitreviews.com/amd-ryzen-7-1800x-1700x-and-1700-processor-review_191753/12

The perf/watt of the 65W TDP R7 1700 is out of this world, just amazing.

R7 1700 = 51,03fps / 124W = 0.4115
Core i7 7700K = 42,32fps / 154W = 0.2748
Core i7 6900K = 57,68fps / 204W = 0.2827

 

[sirmo]

The perf/watt of the 65W TDP R7 1700 is out of this world, just amazing.

R7 1700 = 51,03fps / 124W = 0.4115
Core i7 7700K = 42,32fps / 154W = 0.2748
Core i7 6900K = 57,68fps / 204W = 0.2827

That's going to be money in the bank for AMD. Imagine Naples and Raven Ridge, when they don't have to clock to the moon.

 

[guachi]

You can also quite clearly see what games have really good threading and which ones are subpar. Hint the ones with good threading AMD is really close.

Now that there is finally an affordable processor with more than four cores I hope game studios are pushed to better optimize their games for more cores. We know what well threaded games should look like - BF 1, for example. We just need reviewers to call games out that don't take advantage of it.

 

[Mopetar]

That is the poorest analogy I've ever seen anybody come up with.

Seriously. It happened about 10 years ago and it was called Conroe, and it was every bit as of an amazing jump for Intel as Zen is for AMD.

 

[Absolute0]

Ryzen clearly superior workstation/content creation chip, no doubt there and every sane person must hand it to AMD for making a chip which is massively powerful, energy efficient, and cheap.

That said I want to address the claim "Ryzen was never supposed to be a gaming CPU, so why are all the gamers disappointed?"

The gamers are disappointed because it WAS marketed as a gaming CPU (directly by AMD and then indirectly by pumpers) with built up expectations of parity or superiority in gaming.

"The August demos had Ryzen performing well in both gaming and professional applications."
https://www.forbes.com/sites/patrickmoorhead/2016/12/13/amd-impresses-with-newly-branded-ryzen-cpu-demonstrations-and-product-details/#18e0271f5709

AMD returns to high-end gaming CPUs with Ryzen 7
https://www.engadget.com/2017/02/22/amd-ryzen-7-processors/

Anyone feel like the Ryzen preview was too "Gamer" focused?
https://www.reddit.com/r/AMD_Stock/comments/5ieujn/anyone_feel_like_the_ryzen_preview_was_too_gamer/

AMD will launch its Intel-busting Ryzen PC chips next week
http://venturebeat.com/2017/02/22/amd-will-launch-its-intel-busting-ryzen-pc-chips-next-week/

Maybe they shouldn't claim "performance leadership" off Cinebench alone

Part of this issue wasn't just aiming Ryzen at enthusiast gamers instead of professional workstation users & content creators. It was the fact that WallStreet got high off the $AMD share price and pumped it up with article after article.

The Power of Marketing - AMD's Ryzen Hype Train Hyperloops On
https://www.techpowerup.com/230997/the-power-of-marketing-amds-ryzen-hype-train-hyperloops-on

AMD Ryzen New Benchmark Leaks, World Records, Intel Price Drops And What To Expect On Launch Day
https://www.forbes.com/sites/antonyleather/2017/02/28/amd-ryzen-new-benchmark-leaks-world-records-intel-price-drops-and-what-to-expect-on-launch-day/#7d2a85d225df

AMD Ryzen(TM) 7 Desktop Processors Featuring Record-Breaking Overclocking Performance Available Worldwide Today
http://finance.yahoo.com/news/amd-ryzen-tm-7-desktop-140000157.html

 

[imported_jjj]

A German site tested 8C, 4+2C, 2+2C and 4+0C in 4 games.I don't speak German so can't help with translations.
http://www.pcgameshardware.de/Ryzen-7-1800X-CPU-265804/Tests/Test-Review-1222033/

It doesn't prove much though as with 2+2 the data flow between the 2CCX would be heavier than with 4+4.
It also shows that 8 cores is better than 6 so all in all, don't expect too much from 6 and 4 cores SKUs as more cores is better even if most think the opposite.

 

[Newbian]

Now that there is finally an affordable processor with more than four cores I hope game studios are pushed to better optimize their games for more cores. We know what well threaded games should look like - BF 1, for example. We just need reviewers to call games out that don't take advantage of it.

The problem is they have to build the game around a large amount of hardware and more then 4 cores have been around for awhile on amd's side and this never happened.

Granted it was the fx series so maybe we shouldn't count that.

In either case it would be nice for more games to do this but it would hurt their sales and design costs.

 

[PPB]

That's going to be money in the bank for AMD. Imagine Naples and Raven Ridge, when they don't have to clock to the moon.

Hopefully SenseMI and Pure Power will correct the issues of CPU throttling when the iGP was fully stressed. Because the cpu perf/w is really good at low clocks. Tiled rasterization and HCC can help on Vega to better utilize the iGP (remember CPUs and GPUs arent yet there in regards of knowing WHY they are being pegged at 100% on a certain power state, thus they go to the next one without knowing if the high load is beacuse of the memory subsystem collapsing, this happens a lot on iGP gaming when if memory bandwith isn't there, the iGP bottlenecked tries to boost all the way to P0 thinking it's a core clock problem when in reality the performance will stay lower because of insufficient bandwith) and thus manage power states better, without soffocating the TDP entirely on the GPU and balancing it out with the CPU.

Seriously. It happened about 10 years ago and it was called Conroe, and it was every bit as of an amazing jump for Intel as Zen is for AMD.

Core was more of a disruption because gaming benchmarking was more straightfoward at that time. You either had a fast core or you didnt and it reflected very clearly on framerates. Now with multithreading is more of a mixed bag and you can have cases of Ryzen beating 7700K and other cases where IPC and clockspeed trumps everything.

 

[MaxDepth]

Theory
So AMD is a company different than Intel and nVidia. Both of the other companies have a group that handles graphics (Intel) and SoC (nVidia), but their forte is only one area. AMD however tries to compete against both of these at the same time, right?

So my theory is that these two halves of AMD do talk and they do share. So instead of beating the other companies singularly, they are finding ways to merge their ideas to be better together. That's why the first AMD CPU tests came out using nVidia cards instead of their own. Because some might say, hey, they found a way to produce fake results because their using all their own hardware. And it would be more pronounced if others tested with nVidia and found lesser results (what we see now).

So in the next iteration of their GPUs, will we see really great results of the CPU and GPU working in harmony? I guess I want to believe that to be true.

 

[imported_jjj]

The problem is they have to build the game around a large amount of hardware and more then 4 cores have been around for awhile on amd's side and this never happened.

Granted it was the fx series so maybe we shouldn't count that.

In either case it would be nice for more games to do this but it would hurt their sales and design costs.

The 6900k is already better than the 7700k in gaming, if you actually look into it instead of going with the urban myth built by Intel's marketing. The problem is that the 1800x is too far behind the 6900k- ofc it's minimal at realistic resolutions.

 

[PhonakV30]

http://cdn.overclock.net/a/a3/a3bc7675_Capture1.PNG

Joker posted 720P benchmark and any idea Why does Ryzen have 5300Mb memory but 7700K => 1736mb ?

 

[krumme]

Took 1 hr of bf1 64m amien and sait q. Scar, operations, with a team that really clustered together and pushed on the points. Not a single dip on the counter even the most insane situations. Leages above what my i5 ib 4.2 could muster that was often in the 20ties when a team played like that on those maps. Damn fine subjective experience and the counter agrees. It did what i bought it for and can even handle handbrake fast.
Recommended for bf1 addicts. I am sure this is a fine cpu even in 3 years time. It got tons of horsepower.
Played with r7 1700 on 3.8, 1.33v, 2667c16 ram setting.

Need i say they got whipped !

 

[Agent-47]

So in the next iteration of their GPUs, will we see really great results of the CPU and GPU working in harmony? I guess I want to believe that to be true.

I donot see how though. CPU and GPU talk to each other using specific standard protocols. for AMD CPU and GPU to have a "special bond" they would need a new set of protocols and both drivers and game codes will have to be optimized for those new ports too. At the same time, the standard protocols will also have to be present to allow users the options to choose a Nvidia GPUs as well. this also will result in the making of specialized silicon implementation in both CPU and GPU which may no be utilized if an AMD GPU is used with intel or nvidia GPU is used with AMD.

Whenever AMD tries to push a new standard, they usually make it free for all to use to encourage widespread adoption, live freesysnc or mental, and i see no such push in this case.

 

[Agent-47]

Took 1 hr of bf1 64m amien and sait q. Scar, operations, with a team that really clustered together and pushed on the points. Not a single dip on the counter even the most insane situations. Leages above what my i5 ib 4.2 could muster that was often in the 20ties when a team played like that on those maps. Damn fine subjective experience and the counter agrees. It did what i bought it for and can even handle handbrake fast.
Recommended for bf1 addicts. I am sure this is a fine cpu even in 3 years time. It got tons of horsepower.
Played with r7 1700 on 3.8, 1.33v, 2667c16 ram setting.

Need i say they got whipped !

cool! can i suggest using a frame counter to record GPU usage and CPU usage as well as min and max fps while in multiplayer please.

 

[PPB]

Took 1 hr of bf1 64m amien and sait q. Scar, operations, with a team that really clustered together and pushed on the points. Not a single dip on the counter even the most insane situations. Leages above what my i5 ib 4.2 could muster that was often in the 20ties when a team played like that on those maps. Damn fine subjective experience and the counter agrees. It did what i bought it for and can even handle handbrake fast.
Recommended for bf1 addicts. I am sure this is a fine cpu even in 3 years time. It got tons of horsepower.
Played with r7 1700 on 3.8, 1.33v, 2667c16 ram setting.

Need i say they got whipped !

A German site tested 8C, 4+2C, 2+2C and 4+0C in 4 games.I don't speak German so can't help with translations.
http://www.pcgameshardware.de/Ryzen-7-1800X-CPU-265804/Tests/Test-Review-1222033/

It doesn't prove much though as with 2+2 the data flow between the 2CCX would be heavier than with 4+4.
It also shows that 8 cores is better than 6 so all in all, don't expect too much from 6 and 4 cores SKUs as more cores is better even if most think the opposite.

It can be better in a 1700 vs 1600X scenario. Comparing 1800X and 1600X is absurd as the former costs almost 2x as the latter.

 

[imported_jjj]

With everybody busy with the gaming issues, has anyone noticed this?

https://www.computerbase.de/2017-03/amd-ryzen-1800x-1700x-1700-test/

 

[looncraz]

Zin, I swear upon all things both holy and not, that if AMD fixes these performance issues I will **eat my Ryzen box** and post proof right here in this thread.

https://twitter.com/BitsAndChipsEng/status/837429476237651968

I expect proof to be presented in a timely fashion.

 

[unseenmorbidity]

It can be better in a 1700 vs 1600X scenario. Comparing 1800X and 1600X is absurd as the former costs almost 2x as the latter.

But you can easily oc the 1700 to 1800x levels. I heard ppl were oc'ing further on the stock cooler.

 

[Agent-47]

http://cdn.overclock.net/a/a3/a3bc7675_Capture1.PNG

Joker posted 720P benchmark and any idea Why does Ryzen have 5300Mb memory but 7700K => 1736mb ?

do not forget the GPU usage is lower in intel than it is on AMD while the fps is lower for AMD. lol

also 8 thread at 95% is a bottleneck on Intel, with a total usage of 760%. while the AMD has 70% on average on 16 threads or 1120%.

 

[imported_jjj]

It can be better in a 1700 vs 1600X scenario. Comparing 1800X and 1600X is absurd as the former costs almost 2x as the latter.

Was only noting that as some expect same gaming perf due to similar clocks at stock and the wrong belief that the number of cores doesn't matter

 

[unseenmorbidity]

do not forget the GPU usage is lower in intel than it is on AMD while the fps is lower for AMD. lol

also 8 thread at 95% is a bottleneck on Intel, with a total usage of 760%. while the AMD has 70% on average on 16 threads or 1120%.

Seems like a pointless analysis.

Benchmarks are basically worthless right now, and need redone. Come back in a month.

 

[unseenmorbidity]

yup

 

[imported_jjj]

https://twitter.com/BitsAndChipsEng/status/837429476237651968

I expect proof to be presented in a timely fashion.

Computerbase was noting a huge diff in gaming perf due to BIOS.. https://www.computerbase.de/2017-03/amd-ryzen-1800x-1700x-1700-test/2/
Through Google Translate
" For the AMD Asus Crosshair VI Hero and MSI X370 Xpower Gaming Titanium, the manufacturer recommends at least the use of the BIOS versions 5704 and 117, respectively, and in fact there was a significant gain in performance on the two boards with a double-digit figure To 25 percent. With older BIOS, the Ryzen 7 1800X in games was initially at the level of the Intel Core i7-4770K, which in the editorial memories of bulldozer awakened. The change to the mainboard of Asus and the provision of the BIOS 117 by MSI, however, brought an all-clear.
On the other hand, the two boards still do not lie on top of each other. The main board of MSI is on average three percent, but in extreme cases nine percent. Other BIOS updates should also solve this brake."

If anyone can do a proper translation it would be nice.

 

[unseenmorbidity]

Was only noting that as some expect same gaming perf due to similar clocks at stock and the wrong belief that the number of cores doesn't matter

Except there are a lot of optimizations to be made, as I am sure you are aware by the article you just posted, and this was done at 720p with an enthusiast gpu.

It does not really prove that point.

It would be like saying you have to have a 7700k for gaming, because in these niche situations the 7600k is behind. What matters is if the cpu will be a bottleneck for you. Niche situations are irrelevant.