Official AMD Ryzen Benchmarks, Reviews, Prices, and Discussion

[DrMrLordX]

Not a popular opinion, but I consider Asus an overrated brand. I guess I can't think of a company I'd call a "gold standard" one to replace them in the hierarchy though. They'll get it together on that board eventually, but I'm not sure why everyone is after that particular board now that it sounds like at this moment at least its not a good choice. I guess finding any motherboard is a problem right now.

Sad but true. I went with Asus because they treated the AMD community so well with their Crosshair and Sabretooth boards on AM3+ and the Pro on FM2+. True to form, they did make the (or an) elite board for AM4 . . . but their UEFI squad is really hosing things up now. Give them time, they'll sort out the trouble.

I doubt I'll let my c6H even ship from Amazon though.

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

That seems somewhat accurate actually. There really are problems if you push mem speed too high, so you want to bclk OC instead. Preferably until you push your PCIe devices into 2x mode, since the performance hit is usually not THAT bad. I thought bclk OC actually kept the CPU out of OC mode, though. I could be wrong.

 

[french toast]

Just because I see value in AMD hardware makes me have preference/sympathy/likeness for AMD's products?
I see value in Intel(at this point - not so much, apart from the 35W TDP CPUs, but I am the weird one who builds low-power computers) and Nvidia hardware(GTX 1080 Ti is by far margin Best GPU there has ever been). What does it make me?

The same as me
I too see value in many of intels products and quite openly praise their engineering, i have recommended several of their products as recently as yesterday, ditto Nvida.
But its of obvious who has a 'preference' for a certain company, ive been following this forum for 10 years before i joined, i could rack up a table in 5 minutes of who the sympathizers are on each side, i could then rack up another table of the fanbois on each side just as quick.
Liking a company/product over others is human nature,even if its just 5%. fanboiism is completely different, dont confuse the two.

Im probably 55/45 AMD historically.

 

[guachi]

AdoredTV did a bit of research into this on his latest video using independent benchmarks, there is no evidence that a cpu that performs better at low resolutions will perform better than a rival in future with stronger graphic cards, NONE.
In fact the one metric he found that provides evidence for future gaming performance is THREADS..

RussianSensation harps on this comparing the 2500K and 2600K and how the i7 with more threads just kept pulling farther and farther away.

The same holds true for the 8350 which is now 20% or so faster than the 2500K.

We have evidence from both CPU manufacturers that MOAR cores/threads is a good thing for the future.

Ryzen is good enough now; it'll be even better in the future. The 7700K (or any 4/8 processor) will likely hit a wall sooner rather than later. Though I'm really curious what the R5 or R3 bring to the table. the R7 is "good enough" in games because they priced it low. The R5 may not be enough of a value proposition to be worth it. We'll see.

 

[Bouowmx]

1080p is the new 1366x768; a resolution we should all despise.

But I just got a 1920x1080 144-Hz monitor

 

[Glo.]

The same as me
I too see value in many of intels products and quite openly praise their engineering, i have recommended several of their products as recently as yesterday, ditto Nvida.
But its of obvious who has a 'preference' for a certain company, ive been following this forum for 10 years before i joined, i could rack up a table in 5 minutes of who the sympathizers are on each side, i could then rack up another table of the fanbois on each side just as quick.
Liking a company/product over others is human nature,even if its just 5%. fanboiism is completely different, dont confuse the two.

Im probably 55/45 AMD historically.

Ahem. In timing order: Celeron 700 MHz/Riva TNT2, AMD Sempron/Geforce 6600 GT, Intel Core 2 Duo/GeForce 800GT, i7- 3615QM/GT650M, Core 2 Duo 2.66 GHz/GT330M.


If I ever was a brand fan - it was Nvidia.

Next computer will be most likely based on Ryzen APU, but it requires the hardware to have HBM2 on board for me to buy it. Otherwise I will look for something else.

 

[unseenmorbidity]

Sad but true. I went with Asus because they treated the AMD community so well with their Crosshair and Sabretooth boards on AM3+ and the Pro on FM2+. True to form, they did make the (or an) elite board for AM4 . . . but their UEFI squad is really hosing things up now. Give them time, they'll sort out the trouble.

I doubt I'll let my c6H even ship from Amazon though.



That seems somewhat accurate actually. There really are problems if you push mem speed too high, so you want to bclk OC instead. Preferably until you push your PCIe devices into 2x mode, since the performance hit is usually not THAT bad. I thought bclk OC actually kept the CPU out of OC mode, though. I could be wrong.

According to the post you would push it to 1.0, if you go too far though. PCIE 2 won't matter much,m but I am pretty sure you would see some hefty dropoffs at 1. I asked stilt where that line is, but I don't think he got back to me yet.

EDIT: Apparently stilt doesn't know atm.

 

[hotstocks]

Question:
If I want a Ryzen just for gaming and do what the reviewers say, turn off SMT and overclock it to 3.9GHz, will that make it into a true 8 core (8 thread) chip? Because if it does I would assume there would be less gaming issues and lower heat as a bonus (HT was always hotter and worse at gaming). My point is I would rather have 8 real cores than 16 fake ones that are really 8 pretending to be 16. So for the gamer, 8 real cores would seem to be more than enough and the way to go with Ryzen/SMT off, and then in a year or two when games actually use more than 8 cores/threads turn SMT back on. Seems like we can have the best of both worlds now and later. Only rendering and ripping guys need the fake 16 cores over real 8 cores it seems to me. And 8 real Ryzen cores with SMT off is better than 8 fake cores on an i7-7700 since only 4 are real and the other 4 are hyperthreaded.

 

[DrMrLordX]

But I just got a 1920x1080 144-Hz monitor

Don't feel bad. I'm still running 4:3 for crying out loud.

According to the post you would push it to 1.0, if you go too far though. PCIE 2 won't matter much,m but I am pretty sure you would see some hefty dropoffs at 1. I asked stilt where that line is, but I don't think he got back to me yet.

EDIT: Apparently stilt doesn't know atm.

Yeah, it'll take time and experimentation to figure that out. Plus a board with a proper clock generator. The Gaming 5 doesn't have one, so . . . that board is out.

I noticed the ASRock Killer is now available at the Egg, but I don't think that has one either?

edit: nope, it does not. Plus weaker 45a chokes than the 60a ones on the Taichi.

 

[french toast]

Ahem. In timing order: Celeron 700 MHz/Riva TNT2, AMD Sempron/Geforce 6600 GT, Intel Core 2 Duo/GeForce 800GT, i7- 3615QM/GT650M, Core 2 Duo 2.66 GHz/GT330M.


If I ever was a brand fan - it was Nvidia.

Next computer will be most likely based on Ryzen APU, but it requires the hardware to have HBM2 on board for me to buy it. Otherwise I will look for something else.

I too have had more nvidia products than AMD, doesn't mean anything, just demonstrates your capable of objective purchases and decisions, no fanboi

 

[IEC]

Don't feel bad. I'm still running 4:3 for crying out loud.

On the other hand, I haven't run less than 1200p in over a decade. Presently 4K60 main, 1440p high refresh secondary. After upgrading to 4K main I can't imagine downgrading.

 

[DrMrLordX]

On the other hand, I haven't run less than 1200p in over a decade. Presently 4K60 main, 1440p high refresh secondary. After upgrading to 4K main I can't imagine downgrading.

. . . I don't like widescreen. But this is not the topic for that so, I'll stop before I get started on that.

 

[hotstocks]

The problem is with 4K unless you have a 40" or higher monitor/tv, you can't read anything, it's too small. Sure you can scale it 4X, but that defeats the purpose, you are back at 1080p. I find 1080p is perfect for me to be able to READ text and even text in games need to be read. I prefer two side by side 32" 1080p monitors and I don't sit 10 inches in front of them, more like 3 feet back. Also no modern game will give you over 60 fps in 4K even with a Nvidia 1080 if you turn things to ULTRA. It will take a few more generations of cards to do that.

 

[Pookums]

Hi, I posted something similar on Stilt's technical thread, but as this is more discussed thread I should summarize a bit:

1. UCLK is listed as an option on AMD's own OC Page, but I have yet to hear anything regarding its existence. In advanced core options or elsewhere on any board's bios, can UCLK be modified to approximately double its base speeds(at least up to 2.8 to give it parity to 6900k in this department. Supposedly its only half memory speeds(so even at 3200 memory it would be 1600mhz).

2. If it can be modified, can someone rebench some games and memory latency tests to check improvement.

AND ~

If not, Can anyone take a current GEN intel i7 (preferably a broadwell) and downclock their uncore to the abyssmal speeds of ryzens memory controller, and make a comparison on majincry's drawcall benchmark, on memory latency measurements and by playing games and give a comparison both before and after?

I would really like to get a ryzen processor as I do both gaming and 3d development. Ryzen would be extraordinary, however, as another poster stated above I also purchased a 1080p 144hz monitor. I also have a gtx 1080 waiting to be utilized. Ryzen seems to bottleneck just below that parity around 120fps when ultra settings are applied. The 6900k seemed to max out the 144hz monitor in most of the measured games. Now I will not purchase a 6900k(Too expensive), but I want to make best practice with this monitor and I am hoping that its deficits can be solved. I believe the largest deficits are not SMT or unoptimized code, but rather its abyssmal memory controller speeds that are hindering the ryzen processor atm.

Since no discussion has occured about its memory controller speeds, I'm assuming UCLK is hidden on ryzen. However, If Current intel i7 were shown to have a similar deficit when their uncore was lowered to such a dramatically low level, it might be possible to entice AMD to notify mobo makers to update firmware with proper code to control it and add an option into their bios to unlock it, assuming AMD made their memory controller viable for higher speeds.

 

[inf64]

Wow adoredtv video is simply amazing! It prompted me to follow up on how i5 2500K vs PD 8350 performance developed during last 4 years.

Originally , as adoredtv showed, i5 2500K had around 17%(!!!) percent advantage over 8350 with GTX680 in late 2012 when tested at super low 480p resolution. Logic behind this , like today in Ryzen reviews, is that lower resolution (today's 1080p is yesteryear's 480p) is going to show us how present day CPUs will cope with next gen GPUs in years to come. All makes perfect sense, doesn't it. Lets see.

Fast forward 9 months later, to June 2013, same computerbase did tests, this time with GTX Titan (GK110) (around 35% faster GPU than GTX680), that lead 2500K had over 8350 at 1080p, which was 10.4% has fallen to 8.5% in 2013. Even the difference in 480p super low res. has fallen from 17% to 14% with more powerful GPU! Hmm maybe a fluke result? Lets see what adoredtv found out next.

FFW to 2015 BDW 5775c launch, same computerbase website. The site used the same GPU (Titan GK110) so relative difference between 2500K and 8350 remained the same, no surprise.

FFW to 2017 - the Kaby Lake launch and computerbase's article that covered that. That is 5 years from the original date used in the video for perf. comparison between two popular desktop CPUs: 2500K and 8350. Computerbase used 8370 this time in their article this time as it substituted the original FX8350, but the difference is negligible and is only 2.3% and only when ST is used(base is the same). They used the same GTX Titan GK110. The results? Well, 2500K is 8.2% faster than 8370, same situation as before.

FFW to present day and a week before Ryzen launched. Computerbase updated their gaming test suite and are now using more modern games coupled with an update to more modern GPU in their testbed: GTX Titan X Pascal (the fastest card to date). This should help show that potential 2500K had 5 years back when it had 17% lead over 8350 Piledriver in super low res. 480p gaming - this was specifically done in order to "simulate" how the chip will age with time and how will it handle more modern GPUs and more advanced games.
The results? Today, FX8370 is 10% faster than 2500K in 1080p when tested in latest games, with fastest to date GPU. Even in 720p 8370 is over 10% faster than 2500K!

That is a Bulldozer derivative "8core" chip leading (much) higher IPC 4C/4T Sandy Bridge core in 1080p and 720p gaming, coupled with fastest GPU to date in modern games!
4C/4T is low? Ok, 2600K is just 6% faster than FX8370 in 720P!

Today we can see a similar trend:

4C/8T even though near the top is slowly starting to fall back. In next 2-3 years this trend will continue and all that invested in KL 7700K will see their almost 400$ chip eating dust of Ryzen and BDW-E, SKL-X and Coffey Lake 6C+ chips. This WILL happen and people ought to be aware of this now before they make their purchase decisions.

 

[IEC]

@inf64

What makes that result even more impressive is that Bulldozer was and is a trash CPU from AMD.

It just goes to show that more cores/threads makes a massive difference in the longevity of a CPU, provided the IPC and raw frequency differences aren't too big.

 

[AtenRa]

, but I am the weird one who builds low-power computers) and Nvidia hardware(GTX 1080 Ti is by far margin Best GPU there has ever been). What does it make me?

Infidel

 

[inf64]

@inf64

What makes that result even more impressive is that Bulldozer was and is a trash CPU from AMD.

It just goes to show that more cores/threads makes a massive difference in the longevity of a CPU, provided the IPC and raw frequency differences aren't too big.

The difference today is that (Ry)Zen is high IPC core, unlike Bulldozer derivatives. Zen is massively faster core and it will age even better than Bulldozer since it has no CMT penalty when running more threads, it has 8 fully fledged cores, 60-70% IPC advantage and SMT on top of that. It will be a killer gaming CPU in years to come.

 

[AtenRa]

But I just got a 1920x1080 144-Hz monitor

peasant

Next computer will be most likely based on Ryzen APU, but it requires the hardware to have HBM2 on board for me to buy it. Otherwise I will look for something else.

65W TDP R7 1700 + Vega Nano and you are set to go for eternity

 

[Head1985]

The difference today is that (Ry)Zen is high IPC core, unlike Bulldozer derivatives. Zen is massively faster core and it will age even better than Bulldozer since it has no CMT penalty when running more threads, it has 8 fully fledged cores, 60-70% IPC advantage and SMT on top of that. It will be a killer gaming CPU in years to come.

nope with buged IMC and L3 cache.

 

[ManyThreads]

Hello,

I am new to these boards and I am looking to make a new Photoshop CC workstation. I REALLY want to go with Ryzen but these results made me worried:

https://www.pugetsystems.com/labs/a...n-7-1700X-1800X-Performance-907/#Introduction

Is anyone able to comment on how much the growing pains are likely to have negatively impacted those results? Or do you think that the current issues with Ryzen, once resolved, won't have much impact on Photoshop performance?

I can't get over how much cheaper the Ryzen is than Broadwell-E, and I'd rather not wait until end of August for Skylake-X (and only if it will be priced appropriately...), but some of those performance differences are pretty significant. Of course this is only one test, if anyone has others that are Photoshop related (Specifically photo editing) I'd love to see them.

TIA

 

[french toast]

The difference today is that (Ry)Zen is high IPC core, unlike Bulldozer derivatives. Zen is massively faster core and it will age even better than Bulldozer since it has no CMT penalty when running more threads, it has 8 fully fledged cores, 60-70% IPC advantage and SMT on top of that. It will be a killer gaming CPU in years to come.

Correct, 6900k performance across every scenario for half the price.

 

[zinfamous]

With 4k120 and 8k60 on the horizon, 2560x1440 looks pretty dated. 4k is the new mainstream resolution. 3440x1440 @ 100hz is the only 1440p worth talking about right now and that is going to be outdated in a year or so.

1080p is the new 1366x768; a resolution we should all despise. 1440p is also soon to be a low resolution. Once 4k120 replaces 1440p144 1440p will be dead.

wait, no.

1080p is still, quite easily, mainstream and will likely remain that way for the next 2 years at least? Enthusiast niche gaming preferences will never be mainstream.

I will say that what 4k has going for it right now and in the future, is that displays (60hz) are really quite cheap. But the simple fact that pushing even a "crappy" 4k60hz display currently (and, for many many years to come) requires a minimum of $600-700 investment in GPU will realistically keep it out of mainstream.

As for pushing 4k 120? lol

 

[inf64]

nope with buged IMC and L3 cache.

There are no bugs in IMC and L3. As can be seen above you get ~10% lower fps with Zen vs top of the line intel cores. That is peanuts. The whole platform is rushed which is AMD's bad. They have to work fast and provide OS and firmware updates so the whole platform performs to its fullest potential, that is all. Zen is a killer product.

 

[IEC]

Hello,

I am new to these boards and I am looking to make a new Photoshop CC workstation. I REALLY want to go with Ryzen but these results made me worried:

https://www.pugetsystems.com/labs/a...n-7-1700X-1800X-Performance-907/#Introduction

Is anyone able to comment on how much the growing pains are likely to have negatively impacted those results? Or do you think that the current issues with Ryzen, once resolved, won't have much impact on Photoshop performance?

I can't get over how much cheaper the Ryzen is than Broadwell-E, and I'd rather not wait until end of August for Skylake-X, but some of those performance differences are pretty significant. Of course this is only one test, if anyone has others that are Photoshop related (Specifically photo editing) I'd love to see them.

TIA

If you ONLY use Photoshop CC and you never expect performance/threading to increase, then it seems pretty clear that today the best CPU for your application is the i7-7700K.

However, with a username like ManyThreads I think you should go for the cheapest decent 8C/16T processor, aka the Ryzen R7 1700 @ $329 MSRP. Because... many threads?

 

[CHADBOGA]

With 4k120 and 8k60 on the horizon, 2560x1440 looks pretty dated. 4k is the new mainstream resolution. 3440x1440 @ 100hz is the only 1440p worth talking about right now and that is going to be outdated in a year or so.

1080p is the new 1366x768; a resolution we should all despise. 1440p is also soon to be a low resolution. Once 4k120 replaces 1440p144 1440p will be dead.

Ahhhhhhhh NO.