Official AMD Ryzen Benchmarks, Reviews, Prices, and Discussion

[CatMerc]

HBC != HBM.

HBM would be awesome for performance, but it would have a very small market. An exascale APU, however, may have enough of a market to make it worthwhile.

HBC is specifically designed by AMD to enable bandwidth restricted scenarios to perform MUCH better (+50% avg, +100% min according to AMD... so 20%/40% better ).

XBone does exactly that, but with a strict reliance on software support. If AMD has required software support for the HBC, then they would be stupid to even develop it - a waste of resources (unless it's just for consoles...).

APUs are insanely limited by memory bandwidth. To the point that the lowest APU and highest APU in the same generation perform about the same, despite 33% more resources on the higher end APUs. My A8-7600 with DDR3-2133 CL9 is much better than with DDR3-1600 CL9. It took Hitman Absolution from laughably unplayable to reasonably playable.

HBC helps with capacity restricted situations, not bandwidth restricted. They utilize the high bandwidth to do that, but if you have low bandwidth it won't help.

The 50%/100% result was from testing Vega being artifically restricted to 2GB of VRAM, vs the full capacity of Vega.

With Vega they're switching to a tiled rasterizer, which should help immensly with bandwidth.

 

[PotatoWithEarsOnSide]

Got to question the numeracy skills of TechSpot here. In relation to the 1080p results, the reviewer claims 12% below 6900k. Not sure what math he is using, but neither 111/123 or 114/123, or 123/111 or 123/114 provides the 12% result. (I use both SMT on and off results just to illustrate the point)
That SMT on or off adversely affects both results in any event, without a fix for this, switching between on or off (dependent upon what is best for the game) will result in a figure that is above 114...unless I'm mistaken, right?
Just to clarify, the 1800X provides approx 92-93% of 6900k @ 1080p, or 6900k provides approx 107-108% of 1800X @1080p.

The other thing to note, other than the obvious double posting of 1440p results for the first game, two of the games saw the 1800X yield similar result at 1080p as 1440p (Witchcry 2, ARMA 3). Whilst he talks about optimisation, he misses that the 1080p result is clearly an anomaly.

His conclusion is otherwise very solid.

 

[Whitestar127]

I have logged hours of play time on BF4 and BF1 on my i7-2600k @ 4.5Ghz and my stock-clocked RX 480.

I max everything and run 144hz FreeSync. I *do* artificially limit my max-FPS to 120, so I'll need to make some unrestrained runs and turn off Radeon Chill

I bought a Fury so I could test with it as well.

I will then repeat, with both cards, on both Windows 10 and Windows 7, and report 1% lows, 0.1% lows, averages, and will graph frame times.

I will then do that with my Ryzen 7 1700X at stock clocks... and again overclocked.

LOTS of testing. I best get to it...

Thanks! I have a similar CPU as you (i7-2700K @ 4.4GHz) Will be very interesting to see the results. I thought I would just shamelessly request that you test several other games as well (GTA5, Just Cause 3, Fallout 4, Skyrim SE) if you have any of those.

Admittedly I hadn't been following the Ryzen journey very closely before release. Just heard "competitive with Intel" and "AMD is back". So my slight disappointment in the gaming benches was all on me really. But the prospect of smoother frame rates and devs starting to optimize for Ryzen is exciting. I think I will hold on to my setup for now and see how things pan out during the year.

 

[Glo.]

Yeah, but there are few confusing things about it:
1. I really doubt that this iGPU is Vega based. The Big Die Vega was barely finalized by the time we have heard of this sample.
2. F2 revision? Seriously!?

First engineering Samples of Raven Ridge APUs are dated as far in the past as November 2015. Skeptics, or so they called themselves, believed at that time that even if they are Zen Architecture, which was very unlikely for them, they 100% sure would not use Vega architecture for the GPU.

 

[PotatoWithEarsOnSide]

In the great gaming CPU debate of 2017, between the 7700k, 6900k, and 1800x, the most pertinent question appears to be...


...wait for it...


...what monitor have you got?

 

[Crumpet]

In the great gaming CPU debate of 2017, between the 7700k, 6900k, and 1800x, the most pertinent question appears to be...


...wait for it...


...what monitor have you got?

I've got 2560x1080 144hz, so for me i'm expecting the results to be around the 1440 ones, at which point my old gpu will throttle me until Vega arrives.

The "poor" gaming performance of Ryzen to me is completely blown out of proportion.

I was hoping to be benchmarking by now, but Asus just don't appear to have the motherboards in a high enough quantity, OR they are working on a 2nd revision.

 

[lopri]

RyZEN's performance appears excellent, beating my personal expectation. Admittedly my expectation was quite low in that I thought it would barely be playing a catch-up to Sandy Bridge, which would still not have been a terrible value proposition because 8 core chips from Intel has never been below $500, let alone $330. (not counting ebay transactions) As I see it R7 1700 is the most well-round CPU currently I can find as far as performance, power, and cost are concerned. In a way it is a kind of turbo-charged Jack-of-all-trade with some room to grow. It should do everything with grade A (not A+) now, and it will do everything with grade -A in the future. And by chance if you know how to properly train it you might be able to teach it to behave like an A+ student, which is always a joy to the teacher.

Ordinarily that would be the end of the story, but alas, the product and its platform do not look polished enough for a prime time yet. This is not an unfamiliar ground to old timer like myself who has seen rocky AMD CPU launches of yore, but the technology market sphere has changed quite a bit since then and consumers' expectations are higher and different today, with a lot of value appreciation skewed towards instant gratification. Nonetheless, luckily for AMD, the groundwork they laid out for future products (i.e. Zen's architecture) seems solid - clearly Zen is not a Bulldozer reborn. The question is how quickly AMD will deliver the fixes to the current product/platform (software) and to the future ones (hardware).

P.S. It is enlightening to go back and read some of the BD-E reviews. I will not single out reviews/reviewers, but BD-E's performance characteristics when it debuted was not much different than what RyZen's is now. Yet the shifted focus and manufactured narratives based on selective data are quite different. It is even more absurd because BD-E's prices have been where they were introduced throughout its lifespan. (twice or more the RyZen's prices, and add exorbitant motherboard/memory prices at the time of its launch)

 

[krumme]

Where i got my i5 ib 4.2 below 30fps in bf1 64 man was:
Operations
St scar sector 4-5,
Amiens sector 3-6
And only in situation with a lot of explosion, destruction and man at the same place. So often i could play without any dips even in those maps. If you dont have people clustering together with making loads of explotions and destruction it will never happen or even get close to it. My guess is operations in those sectors and maps gives those situations.
I never experienced problems in conquest for that reason i guess but i have only played it little.
A typical situation forcing the cpu to extreme low fps is if 20 man and 2 tanks try to take down a train and the typical 30 men having their fight nearby.
I dont think its background processes because its always under severe load (looks like a physics issue to me - and looks to happen when there is lots of destruction). I only play assault btw - perhaps that add to how often it happens because i always go to take the tank/train whatever as first priority. So i always end where there is most destruction.

So in reality i dont think neither 0.1 or 1% catches those problems. The problem is how to replicate it "all snipers change now please" .
Now its only interesting for those that play BF1 and only if you play on those situations where there is huge cluster situations with men and material that fighting in sectors in operations gives.

I even managed to get my stock 1700 to a 58 fps dip. 3800 is 100%.

I dont know if anything is wrong with my setup, but if people dont play in those situations i am sure they get something like +100fps all the time.

 

[Glo.]

What is responsible for high minimum framerates in Ryzen CPUs? Number of cores?

Its interesting in the context of 6 and 4 Core CPUs.

 

[CHADBOGA]

An In-Depth Look at Ryzen's Gaming Performance: 16 Games Played at 1080p & 1440p

http://www.techspot.com/review/1348-amd-ryzen-gaming-performance/

Leaving Ryzen out of it for now, can we at least put an end to those silly people claiming the 6900K is a better gaming chip than the 7700K.

 

[imported_jjj]

Found this Google Trends ss, Ryzen vs aliens and thought it's pretty hilarious that Ryzen is more popular.

 

[guachi]

What the techspot article of 16 games and some online videos I've seen show is:

Your GPU matters vastly more than your CPU, especially if you own a 60Hz monitor.

You can game just fine even with a lowly FX8300 series chip if you turn the settings down or if you own anything slower than a 1080 GPU. I'd be shocked if there was much difference in any game at 1080 resolution on a 480 between an FX-8370 and a 7700K.

If you own a 1080 GPU, why are you gaming at 1920x1080 resolution? Why test at this resolution with this GPU? Test a 1070 at this resolution, please.

A 1700 that you can OC to 3.7 GHz with the supplied cooler is a steal at $330. It's as future proof as you can get for a CPU at that price and it crushes anything else in performance/dollar for any other CPU tasks.

EDIT: If you are curious, the 1800X has an overall score of 91% vs. the 7700K. It's basically equal to the $240 6600K, though I can't see recommending the 6600K over something like the 1700 at $330. It'd just be the i5-2500K vs. i7-2600K all over again where the 2500K was a great deal until it wasn't and eventually even the FX-8350 surpassed it.

 

[imported_jjj]

Leaving Ryzen out of it for now, can we at least put an end to those silly people claiming the 6900K is a better gaming chip than the 7700K.

He tests with a Titan X Pascal at 1080p and 1440p, that's low res and in just 16 games.
Do note that the 6900k doesn't have Turbo 3 enabled in the review and that would be an extra 8% ST clocks.
And next time, before insulting people, go do some research.

 

[Crumpet]

I'm thinking if the R5 can produce the SAME gaming results as the R7 series, but for $260 vs $350 for the i7 7700k, with vastly cheaper motherboards, for the same price you could stick a significantly better GPU in the AMD build...

It will be hard for me to recommend the i7 7700k for someone unless they are on an unlimited budget, as the R5's should provide some serious bang for the buck.

 

[Malogeek]

I'm thinking if the R5 can produce the SAME gaming results as the R7 series, but for $260 vs $350 for the i7 7700k, with vastly cheaper motherboards, for the same price you could stick a significantly better GPU in the AMD build...

It will be hard for me to recommend the i7 7700k for someone unless they are on an unlimited budget, as the R5's should provide some serious bang for the buck.

Yeah I don't think gaming performance will be very different with the R5 series and clock speeds shouldn't be an issue, if anything higher than the R7 series? It should be a very compelling CPU. I always thought the R5 will be the sweet spot but I simply couldn't wait for it to be released.

 

[zinfamous]

I'm thinking if the R5 can produce the SAME gaming results as the R7 series, but for $260 vs $350 for the i7 7700k, with vastly cheaper motherboards, for the same price you could stick a significantly better GPU in the AMD build...

It will be hard for me to recommend the i7 7700k for someone unless they are on an unlimited budget, as the R5's should provide some serious bang for the buck.

yeah, these are the performance numbers I am waiting for after maturity and when the 1600x is available, vs the 1700. If the 1600x is at the relatively same spot with gaming performance (if not better for some reason), and marginally worse at mt tasks, then I will pump those savings into a more spendy GPU. Man, I sure do wish today was ~July. But it's all good...I've got a significant amount of expenses throttling my budget starting the next month, so I'll need that 2-3 month time gap to adjust and get a better idea on what I can comfortably afford after that time (a new 32-34" 1440p Async/100hz+ display will be part of that, so....)

 

[GaiaHunter]

Ryzen is a good product.
Obviously going forward we hope a Zen+ will have higher IPC and more OC headroom.

Additionally for gaming for the last 10 years Intel was the only player in town, so it is no surprise that games will be optimized for Intel architecture.

Hopefully because Ryzen is a good product and R5 should make pretty decent gaming boxes for a lower price than i7 6700/7700, games will support Zen architecture better and we will see more comparable results.

And if you are currently running a GTX1600 or a RX480 @1080p the game result differences will be way smaller between Ryzen and Intel products.

 

[unseenmorbidity]

Apparently 3200 is borked on Ryzen.

 

[Crumpet]

Yeah I don't think gaming performance will be very different with the R5 series and clock speeds shouldn't be an issue, if anything higher than the R7 series? It should be a very compelling CPU. I always thought the R5 will be the sweet spot but I simply couldn't wait for it to be released.

Same, I WAS going to get an R5 but due to R7's releasing that much earlier I took the jump.

 

[jaymc]

I hear that Ryzen is running like a swiss watch on Linux...

Also I hear the patches are already out for scheduling..

It would be very interesting to see some gaming stats running on Linux

edit spelling

 

[unseenmorbidity]

Overclocking the base clock (BCLK) on AM4 platform is possible, however generally not recommended. This is due to its frequency relations with other interfaces, such as the PCIe. Unlike with Intel's more recent CPUs, there is no asynchronous mode (straps / gears) available, which would allow stepping down the PCIe frequency at certain intervals. The PCIe frequency relation is fixed and therefore it increases at the same rate with the BCLK. Gen. 3 operation can generally be sustained up to ~107MHz frequency and higher speeds will usually require forcing the links to either Gen. 2 or to Gen. 1 modes.

motherboards have the external BCLK generator (you can't change BCLK without it).
Also I heard BCLK clock = PCIE clock (AM4 platform uses 100MHz BCLK as opposed to 200Mhz of previous AMD platforms, PCIE speed wasn't linked to BCLK in those) which can lead to many problems, from simple instability to data corruption.
Higher memory ratios are only available on a few motherboards and even on those they're very dodgy and unstable. Also ryzen runs very loose memory subtimings at higher ratios, defeating the point of higher frequency memory and sometimes causing instability not because the subtimings are too tight but quite frankly because they're too loose.
Overall to achieve these kits' advertised speeds you will need a high end motherboard with external clock generator (only ASUS C6H is validated by G.Skill at the moment - the board that is currently plagued by a BIOS issue that can randomly brick it), a ryzen sample that works fine at nearly 120MHz BCLK (I heard not everyone was able to reach that value, someone somewhere mentioned having issues with going past 107MHz, unfortunately I can't remember who) which also means that you need a good ryzen sample in general because manipulating BCLK or multiplier disables boost and XFR so unless you can achieve the boost/XFR speed on all cores you're losing single threaded performance.

Quotes are from random people on reddit. I cannot confirm any of this.

 

[CentroX]

My asus ch6 hero got bricked. I started this huge thread here

https://www.reddit.com/r/Amd/comments/5xn51g/my_asus_crosshair_vi_hero_got_bricked/

 

[PotatoWithEarsOnSide]

Pretty sure those are quotes taken from The Stilt's review, given that he seems to be the only person on the planet that has actually done a technical review of Ryzen.

 

[unseenmorbidity]

My asus ch6 hero got bricked. I started this huge thread here

https://www.reddit.com/r/Amd/comments/5xn51g/my_asus_crosshair_vi_hero_got_bricked/

Looks like the nazi mod didn't like your critical review of AMDs supposed flagship motherboard.

 

[CatMerc]

I took Hardware Unboxed video, and removed any result that is obviously misbehaving one way or another (Which would result in a very very large delta from 6900K). I ignored large deltas from 7700K (as in kept them in), as that is perfectly explainable by higher clockspeed and IPC. I then from these games took the highest Ryzen result, whether it is SMT on or SMT off, since I assume after a few patches here and there to Windows and games, the CPU will perform at least as fast as the best result there.

Games Removed:
Hitman
Watch Dogs 2

I also made two separate averages with and without Deus Ex, as the minimum appears oddly out of line compared to the averages.

With Deus Ex (Min/Avg):
Ryzen: 94/118 (100%/100%)
7700K: 110/137 (117%/116%)
7600K: 90/123 (95%/104%)
6900K: 98/124 (104%/105%)

Without Deus Ex (Min/Avg):
Ryzen: 97/120 (100%/100%)
7700K: 112/139 (115%/116%)
7600K: 92/125 (94%/104%)
6900K: 100/126 (103%/105%)

So, ignoring outliers and accounting for SMT regression, 6900K has 5% higher averages and 3% higher minimums.
Not bad at all for buggy rushed Microcode with poor scheduling from Windows.