Official AMD Ryzen Benchmarks, Reviews, Prices, and Discussion

[Crumpet]

It's basically to the point that I just can't for the life of me even acknowledge a review that doesn't at least try to include a Radeon. I was going to wait for Vega and then decide which between the 1080 and Vega to be my system video card for the next couple years. But if Vega comes in overpriced and slow I'll just get a 570 or 580 to carry me over till the next big Radeon release. I am just not going to participate in this madness of if and when Nvidia might get around to it.

I bought a freesync monitor so i'm going Vega regardless. Though to be honest I expect 1080+ performance for 500ish, which is good enough for me. I can't see it being slower than that to be honest.

I can't be the only one wondering if Nvidia has some kind of deal with Intel.

Is it? It's not like AMD sent Nvidia a Ryzen reviewer kit with words: "Hey folks, mind optimizing your driver for our CPU?".

Well they're sending them out to a whole bunch of game devs, why wouldn't they send one to Nvidia? Unless you have evidence to backup your BOLD claim.

 

[Despoiler]

I think the 6/12 parts are a great value. I am also wondering if Nvidia was ignorant to the fact that their drivers were crippling the hell out of Ryzen chips. My opinion is that they knew about it and could have easily fixed it but decided it wasn't their responsibility to do so and if their rival gets injured as a result of Ryzen's resulting image problems regarding gaming performance, then they figured "oh well"?
This is a fantastic and very sleazy way for Nvidia to dissuade people from buying Ryzen. They know they have the only viable GPU's for gaming these days and if you want an NVidia GPU, you are likely to reconsider a Ryzen purchase, and any money AMD doesn't get for Ryzen sales also hurts their GPU division.

Nvidia claims they didn't get a Ryzen test system until launch. I wouldn't put anything past them though given the business strategies they choose to employ.

 

[Crumpet]

http://www.gamersnexus.net/hwreviews/2875-amd-r5-1600x-1500x-review-fading-i5-argument?showall=1

Reading through this, effectively giving GN a second chance, and all they are talking about is Max FPS, not mins or averages.

This guy is a hack. No more page views for him.

 

[nathanddrews]

Looking at some of the AT review gaming graphs, in particular the time spent under 60fps and under 30fps, I think I can see why some people claim Ryzen to be "smoother". Whereas Intel dominates the 120fps+ arena, Intel also seems to dip down below 30fps and 60fps more frequently. Ryzen seems to be more consistent, albeit it "slower" overall for games. I think we knew this already, but it is much more apparent with the i5 CPUs than the i7 CPUs.

 

[kush120]

This is what I've been talking about. If you benchmark just the gaming running like most reviewers these days, the Intel CPU would be winning this test.
But you add streaming, or anything else for that matter that happens during normal everyday usage. The Ryzen just pulls ahead.

I own both a 1700 and 6600k and its very pronounced as soon as you start doing anything else while gaming. Even just watching HD video, talking on Discord, having multiple browsers, and other programs open it starts to pull away for Ryzen.

 

[Topweasel]

For w/e reason i can tell you were not planning on getting nV either way, the CPU performance debacle or not.

You say that because I have been going on about how I am pretty sure Nvidia's DX12 driver is boned for core scaling. But that's my point Nvidia basically agrees but the best we saw was one tweet saying "hey guys we haven't had Ryzen for very long" that was 2 weeks ago. Now 6 weeks after release and from a company that has been able to do same week beta updates to fix game bugs in the past. I am getting more and more certain they aren't going to fix this because it will eat away some of the top end DX12 performance on the 7700 (the only CPU to see consistent performance on DX12 and little to no CPU overhead drops from DX11).

That said my first video card was TNT2, I had a Geforce 256, Geforce 2, Geforce 4 and Geforce 4mx, A 7800GTX, 7900GT, 8800GT, and several others. I have been more AMD centric lately as I haven't upgraded my video card as often. My Phenom X4 used a 5870 for 4 years (best price/perf at the time) and a 7970 in my 3930k system that I am borrowing for my Ryzen setup till I figured out which card to get. I got the 7970 because there was still a question on when Kepler would hit, whether it fixed Fermi's issues power wise, and the idea of the 104 being the mid range "chip" bothered me for the premium prices. In the end I wasn't heart broken with how it turned out. The Titan was never an option, the 680 was marginally better, and Mantle was an awesome idea that basically got MS to announce that they were going to buckle down and get DX12 right.

Unlike my opinions of Intel. I really don't mind Nvidia. Infact some of my best memories gaming and even PC building were from working with Nvidia systems. They could be a better company for Consumers. But its not like AMD didn't do the same once they built up clout with superior chips like (well ATI then)1900xtx and their pricing on the 6900 and 7900 series. Or the pricing of the Athlon64 and Athlon64 x2 (and their FX counterparts), when they had the CPU lead. But I am not a 18 year old kid only worried about gaming anymore, Ryzen was a happy choice, but a choice nonetheless about productivity. If that choice is going to continue to be treat like a redheaded stepchild by Nvidia, then it makes the choice easier, even its slower at least I am not going to get wildly different experiences at random. Really Nvidia still has probably another month to fix it's drivers.

As for reviewers not showing a Radeon means what we have already seen randomly skewed tests that make Ryzen look worse than the competition where we see a more balanced look at CPU performance on Radeon cards. We need both because A.) It's nice to see what the fastest will get you and B.) seeing issue and measuring helps people realize what is happening and increases visibility that will help push for a fix.

 

[lolfail9001]

Well they're sending them out to a whole bunch of game devs, why wouldn't they send one to Nvidia? Unless you have evidence to backup your BOLD claim.

In case you missed the memo, i understand you mean more than just those few game devs AMD are partnered with and i do mean the pre-launch reviewer kit, because you are not going to optimize a driver for a new arch altogether in a month that nV had so far.

But you add streaming, or anything else for that matter that happens during normal everyday usage. The Ryzen just pulls ahead.

CPU Streaming is nothing like normal everyday usage for about everyone running games, ya know. I agree that as long as you are doing something heavy with CPU while gaming, Ryzen is a no-brainer over Intel for the money. How many do, though?

You say that because I have been going on about how I am pretty sure Nvidia's DX12 driver is boned for core scaling.

No, i forgot about that, i had your mention of purchasing an rx480 rebrand instead of rx480 should Vega flop in mind.

Now 6 weeks after release and from a company that has been able to do same week beta updates to fix game bugs in the past.

They have the copy of games they fix months in advance for that. And game code is a naturally more transparent than CPU behavior.

 

[coercitiv]

Really Nvidia still has probably another month to fix it's drivers.

If there's anything to be fixed, they will definitively do it before the Vega launch, so everybody can relax.

 

[beginner99]

I think the 6/12 parts are a great value. I am also wondering if Nvidia was ignorant to the fact that their drivers were crippling the hell out of Ryzen chips. My opinion is that they knew about it and could have easily fixed it but decided it wasn't their responsibility to do so and if their rival gets injured as a result of Ryzen's resulting image problems regarding gaming performance, then they figured "oh well"?
This is a fantastic and very sleazy way for Nvidia to dissuade people from buying Ryzen. They know they have the only viable GPU's for gaming these days and if you want an NVidia GPU, you are likely to reconsider a Ryzen purchase, and any money AMD doesn't get for Ryzen sales also hurts their GPU division.

Or people go out there build a ryzen system adn then see that an NV card sucks and choose AMD instead.

 

[unseenmorbidity]

Is it? It's not like AMD sent Nvidia a Ryzen reviewer kit with words: "Hey folks, mind optimizing your driver for our CPU?"..

AMD could have offered them engineering samples, or they could have not. Nvidia could have refused the engineering samples, or they could have accepted them, and did nothing. Also, Nviadia might have gotten samples, and they are working on it now.

All are possible scenarios. Do you have some insight to share as to which is true?

 

[SPBHM]

hopefully Nvidia will patch their drivers soon, that result on Rocket league is super bizarre;

other than that, the 1600 is the champion right now, you can still justify the 7600K with OC for many cases (likely for games like Arma3, emulators and so on), but overall the 1600 is far better, it's clear that right now many games are taking good advantage from more than 4 threads, and the single thread performance of Ryzen is not terrible anyway, still Kaby lake at 4.8GHz is a lot faster for that

 

[thepaleobiker]

Reading through this, effectively giving GN a second chance, and all they are talking about is Max FPS, not mins or averages.

This guy is a hack. No more page views for him.

I watched the youtube video (liberally fast-forwarding when required) and GN seemed to have detailed metrics for games, including Ave FPS, Min 1% numbers as well as 0.1 % numbers.... Although R5s on average are about 5%-10% below i5 Kaby Lake equivalent counterparts in average FPSs', they are atleast 10-30% higher on minimum 0.1% FPS and 1% FPS metrics in many games.

Metro : Last Light is an Extreme EXample

BF:1 is a more typical example

Their Summary :-

Conclusion: i5 Hangs On with Fading Grasp
There’s no argument that, at the price, Ryzen is the best price competitor for render workloads if rendering on the CPU – though GPU-accelerated rendering does still serve as an equalizer, for people who use compatible workloads (see: Premiere w/ CUDA on i5-7600K, 6900K, & 1800X). If CPU rendering is your thing, Ryzen 5 is well ahead of same-priced i5 CPUs.

For gaming, AMD ties same-priced Intel i5 CPUs in some games – like Watch Dogs 2 before OC – and is 7-15% behind in other games (7-10%, generally). AMD has closed the gap in a significant way here, better than they did with R7 versus i7, and offers an even stronger argument for users who do legitimately plan to do some content creation alongside gaming. With regard to frametimes, AMD’s R5 series is equal in most worst cases, or well ahead in best cases. Although the extra threads help over an i5 CPU, the R7’s extra threads – Watch Dogs notwithstanding – do not generally provide much of an advantage.

Regards,
Vish

 

[CatMerc]

In honor of the R5 series launch, I made this:

 

[mohit9206]

Six core 12 thread 1600 is a great value for money and the most futureproof cpu for the money.

 

[Crumpet]

I watched the youtube video (liberally fast-forwarding when required) and GN seemed to have detailed metrics for games, including Ave FPS, Min 1% numbers as well as 0.1 % numbers.... Although R5s on average are about 5%-10% below i5 Kaby Lake equivalent counterparts in average FPSs', they are atleast 10-30% higher on minimum 0.1% FPS and 1% FPS metrics in many games. ~etcetc

Good post, thanks for sharing.

Seems as I figured. Unless all you care about is the biggest FPS digit (and can ignore the crippling stuttering lows) the R5's are cleaning house overall.

 

[unseenmorbidity]

Good post, thanks for sharing.

Seems as I figured. Unless all you care about is the biggest FPS digit (and can ignore the crippling stuttering lows) the R5's are cleaning house overall.

Gamer's Nexus relies on flawed benchmarks, rather than actual gameplay. The 1600x would absolutely slaughter an i5 in real bf1 gameplay.

 

[Crumpet]

Gamer's Nexus relies on flawed benchmarks, rather than actual gameplay. The 1600x would absolutely slaughter an i5 in real bf1 gameplay.

Yeah, like I said I wanted to give him a second chance. I gave Jayz2Cents a second chance and to be fair to him, his latter Ryzen reviews have actually been fair. (not seen anything from him re: R5 yet though), so fair's fair i'd give GN a second chance.

But no, still as ignorant, still full of crap.

 

[Malogeek]

I own both a 1700 and 6600k and its very pronounced as soon as you start doing anything else while gaming. Even just watching HD video, talking on Discord, having multiple browsers, and other programs open it starts to pull away for Ryzen.

I can play Battlefield 1 multiplayer whilst processing a h265 video in Handbrake in the background. Ryzen power

 

[Topweasel]

No, i forgot about that, i had your mention of purchasing an rx480 rebrand instead of rx480 should Vega flop in mind.

That is why I floated that. I am not going to pay a premium to dicked around. That includes AMD I am not going to buy an overpriced "highend" GPU unless it performs like a high end card. Just not being dicked around with the drivers isn't enough for me to do so.

So 1080, generally fast, but is an arch dead end for DX12 and Vulcan, on top of that drivers that penalize my CPU choice and no impetus to fix. Which isn't a good sign if I plan on using the card for a long time.
Vega, If expensive and slow, then even if better for DX12 isn't worth spending money on.
1070 Reasonably priced but still drivers dicking people around.
1060/580. Cheap almost disposable price, 1060 still arch dead end and redheaded stepchild platform support. 580, while the arch is at a dead end its better than all other non Vega chips in DX12.

The choice becomes really easy. A 580 isn't priced high enough I won't mind to upgrade it in a year or two. I'll get good platform support. And it will age better with increased DX12 and Vulcan development. I'd rather get a $400-$600 I won't feel the need to upgrade. But I have enough systems that I can always move the 580 later.

 

[unseenmorbidity]

Yeah, like I said I wanted to give him a second chance. I gave Jayz2Cents a second chance and to be fair to him, his latter Ryzen reviews have actually been fair. (not seen anything from him re: R5 yet though), so fair's fair i'd give GN a second chance.

But no, still as ignorant, still full of crap.

I agree. I was actually fairly impressed that he spoke out against Nvidia about their recent Titanxppp release. He is on nvidia's short list for titan caliber cards.

 

[AMDisTheBEST]

The fact that ryzen r5 only overtakes the Intel i5 as resolution is crank up to 4k tells me there is defiantly something fishy about how the gpu works with the CPU.....

 

[krumme]

Looking at some of the AT review gaming graphs, in particular the time spent under 60fps and under 30fps, I think I can see why some people claim Ryzen to be "smoother". Whereas Intel dominates the 120fps+ arena, Intel also seems to dip down below 30fps and 60fps more frequently. Ryzen seems to be more consistent, albeit it "slower" overall for games. I think we knew this already, but it is much more apparent with the i5 CPUs than the i7 CPUs.

Gaming experience have never been about avg framerates but about 99% and time under 30fps. Those moments define if you kill the other guy or he kills you.
When i look at AT new bm its aparent to me Ryzen is not "smoother" - i dont care - its just flat out faster for gaming. And the numbers scream it out loud.

And NV showoff....ouch. lol. They better improve or have a good explanation because that looks like the most reactionary and destructive behavior. But lets see if it doesnt change. Lets hope so.
Man i was entertained.

 

[nathanddrews]

The fact that ryzen r5 only overtakes the Intel i5 as resolution is crank up to 4k tells me there is defiantly something fishy about how the gpu works with the CPU.....

I just assumed that was cache-related. The i5 models are cut down more (6MB) than the R5 models (16MB).

 

[sm625]

Reading through this, effectively giving GN a second chance, and all they are talking about is Max FPS, not mins or averages.

This guy is a hack. No more page views for him.

One thing for sure, this is an example of how NOT to format a chart axis:

I honestly dont see how something like that gets past even the most basic proofing process.

 

[guachi]

I think (and I've said this before) that reviewers should dump average fps and give us median instead. Yes, it will likely be close but it gives you something useful. Half your time your fps will be above this and half the time your fps will be below this.

In addition, we should continue to get 1% and 0.1%, though .1% can the thrown off easily enough I'd run the tests several times to eliminate outliers, if possible.

Down with average fps!!!!

That being said, the 1600 at $220 with cooler and a B350 motherboard is an insane value allowing a mild OC out of the box for cheap.

Pair it with a 480 (or 580) that won't screw you over in games because of drivers that can't handle more than 4 cores or DX12 and you have a fantastic gaming machine at 1080. Oh, and pair it with a Freesync monitor that doesn't have a stupid tax on it. You'll have a smooth gaming experience (either because of Freesync or MOAR CORES) and you'll have enough cores/threads to still do something else on your machine while you game.

If I sound like an AMD shill, I'm OK with that. Because we now have enough evidence that the AMD ecosystem can provide a great gaming + other stuff experience on your computer for a mid-range gamer who doesn't have money to spend (or just doesn't want to spend the money)

For a higher end gamer... well, Vega better not suck. If it does, AMD has a stand-up double. If it doesn't suck, they have a home run, IMO.