Discussion AMD Ryzen MegaThread! FAQ and Resources

[jaymc]

Dota 2 before and after Ryzen patch:

https://hardforum.com/threads/dota-2-update-increases-performance-for-ryzen.1927957/

The OP claims the patch improved performance by 23% on RyZen. The i7 7700k was a huge 50--60% ahead the R7 1800x before the patch

http://www.phoronix.com/scan.php?page=article&item=nvidia-1080ti-ryzen&num=2

But the problem is solved... ana 23% improvement is no small gain... So patches can an will make a big difference it seem's... Doesn't really matter if it catchs kaby lake... As the difference is impercievable anyway while playing that is...

People were complaining that frame rates were abnormally lower in certain titles... This proves that patches will indeed fix this for the games that are not showing expected results !!...

It was silky smooth experiance anyway. Once fps is certain amount it's unoticeable. The hole grip in the first place was not weather it was faster or as fast as kabylake (it was never targetted at kaby lake or compared to it either, it was targetted at 6900k... )...
You cannot notice anyway... as long the cpu is fast enough for your gpu thats all that really matters...
look at the performance in windows an multitreaded app's & video encoding..
0nce it's fast enough to keep the gpu busy happy days it can run games end of. I tell you what though it's a beast in other departments... and an absolute bargain in comparison to 6900k

 

[Nope 1151]

Has anyone seen the rumors of Ryzen 5 selling early?
I've seen them in some (South American?) shops on Chinese forums.

Also, I thought this would be funny: https://www.newegg.com/Product/Product.aspx?Item=N82E16834234538

 

[Redneck5439]

It looks like bios updates, RAM optimizations, and microcode updates have all helped Ryzen overclock in only 3 weeks.

https://www.youtube.com/watch?v=I71ov8iiDP8

It will be interesting to see how far AMD will be able to push Ryzen with further updates and optimizations. They have made good progress in only three weeks, 4.1Ghz on all 8 cores with RAM at 3000Mhz is pretty good for a brand new architecture. Very impressive as only 3 weeks ago Ryzen couldn't be pushed past 3.9Ghz and 2133Mhz on the RAM.

 

[JoseTheBestOle]

With 3600 RAM the gaming gap with the 7700k is closed.

So now is pointless to say "7700k if you game only"
But imho it's still ok to say "A PS4 if you game only"

There you go:

https://www.youtube.com/watch?v=RZS2XHcQdqA

 

[jdwii]

With 3600 RAM the gaming gap with the 7700k is closed.

So now is pointless to say "7700k if you game only"
But imho it's still ok to say "A PS4 if you game only"

There you go:

https://www.youtube.com/watch?v=RZS2XHcQdqA

Again terrible bench marking ram frequency should be the same Intel chips scale quite well to memory frequency.

 

[con635]

With 3600 RAM the gaming gap with the 7700k is closed.

So now is pointless to say "7700k if you game only"
But imho it's still ok to say "A PS4 if you game only"

There you go:

https://www.youtube.com/watch?v=RZS2XHcQdqA

The 4 and 6 core ryzens will be very interesting in reviews. It will soon be hard to recommend intel without price drops.

 

[Redneck5439]

With 3600 RAM the gaming gap with the 7700k is closed.

So now is pointless to say "7700k if you game only"
But imho it's still ok to say "A PS4 if you game only"

There you go:

https://www.youtube.com/watch?v=RZS2XHcQdqA

Great video, thanks for sharing. As bios updates, motherboard optimizations and microcode tweaks mature its going to be interesting to see just how far this new Ryzen platform can be pushed. AMD does have a winner on their hands this time and its nice to see how continuing updates are improving performance. Intel may always be "better" gaming than Ryzen, but as long as AMD keeps things very close then that performance "increase" becomes more and more mundane as there is no way in real world playing you would ever even notice the difference between the two platforms. Couple that with Ryzen's amazing workstation prowess and AMD has a very heavy hitter to compete with Intel.

 

[jaymc]

With 3600 RAM the gaming gap with the 7700k is closed.

So now is pointless to say "7700k if you game only"
But imho it's still ok to say "A PS4 if you game only"

There you go:

https://www.youtube.com/watch?v=RZS2XHcQdqA

The 4 and 6 core ryzens will be very interesting in reviews. It will soon be hard to recommend intel without price drops.

An the guy mentions the rumour that Microsoft did a silent update for Win10 & an how Ryzen it handles treads on the 16th....

 

[TechyInAZ]

It's interesting that Memory frequency is the key to overclocking Ryzen. If you guy's missed the latest WAN show...According to several news sites, having a higher frequency on the RAM increases CPU overclockability. So memory performance is incredibly crucial for Ryzen to perform well.

 

[Rookie_MIB]

It's interesting that Memory frequency is the key to overclocking Ryzen. If you guy's missed the latest WAN show...According to several news sites, having a higher frequency on the RAM increases CPU overclockability. So memory performance is incredibly crucial for Ryzen to perform well.

Well, I'll take a whack at memory timings (loosen them up) and see if I can't get faster RAM speed working at home and see how that works out. As it sits right now, my 1700 is humming along quite happily at 3.7 on all cores at a very reasonable idle/load temp and voltage with the stock cooler.

 

[aldaia]

That is the top SKU. I would expect 12 core models as well, from combining a pair of six-core dies.

Apparently the 12C24T has already been spotted.
https://videocardz.com/67649/amd-ryzen-cpu-with-12-cores-and-24-threads-spotted

Some sources claim it will be the 1900X, and 1950X will be the 16C32T

 

[8350rocks]

It's interesting that Memory frequency is the key to overclocking Ryzen. If you guy's missed the latest WAN show...According to several news sites, having a higher frequency on the RAM increases CPU overclockability. So memory performance is incredibly crucial for Ryzen to perform well.

Well, I'll take a whack at memory timings (loosen them up) and see if I can't get faster RAM speed working at home and see how that works out. As it sits right now, my 1700 is humming along quite happily at 3.7 on all cores at a very reasonable idle/load temp and voltage with the stock cooler.

PROTIP: For anyone trying to get faster memory working:

1. Higher memory frequencies from GSkill were achieved by using a MB with external clock gen to run the tighter timings. (i.e. 2933 timings @ ~118 MHz bclk for 3466). Currently those boards include the X370 Taichi and Fatal1ty from Asrock, the CH VI Hero from Asus, and Gaming K7 Aorus from Gigabyte.

2. If using external clock generator to overclock and boost RAM frequency, avoid going past about 120-125 on bclk as it throws off the PCIE3 timings, and the chip defaults to PCIE2.

3. Check your RAM to ensure it uses Samsung "B" die memory modules. These only come in sticks that are 8 GB or 16 GB per stick.

4. Ensure you are running newest BIOS, etc. for best results.

 

[vvacenovski]

Out of "budget" X370's, would you recommend Asus PRIME-pro or Asrock K4?

 

[juanrga]

Dota 2 before and after Ryzen patch:

https://hardforum.com/threads/dota-2-update-increases-performance-for-ryzen.1927957/

The OP claims the patch improved performance by 23% on RyZen. The i7 7700k was a huge 50--60% ahead the R7 1800x before the patch

http://www.phoronix.com/scan.php?page=article&item=nvidia-1080ti-ryzen&num=2

But the problem is solved... ana 23% improvement is no small gain... So patches can an will make a big difference it seem's... Doesn't really matter if it catchs kaby lake... As the difference is impercievable anyway while playing that is...

The problem is not solved. What they did was to take the worst game and patch it. No one would expect 23% improvements for any patched game and no one would expect all games to be patched.

 

[juanrga]

That is the top SKU. I would expect 12 core models as well, from combining a pair of six-core dies.

Apparently the 12C24T has already been spotted.
https://videocardz.com/67649/amd-ryzen-cpu-with-12-cores-and-24-threads-spotted

Yes. The reported amount of L3 cache looks wrong. It would be 4 x 8MB.

I guess A9 = 150W.

 

[IgObYaGo]

Hey all, I just ran my first bench test upon my first start with my Ryzen. I am using my stock AMD cooler from my old Phenom II x6 1035T (I have a Corsair H100i v2 closed loop water cooler, but the AM4 bracket isn't out yet). Here is my first test score on Single Thread and Multi Thread: ST= 2165 MT= 18720 on the CPU-Z test.

 

[-Fran-]

Dota 2 before and after Ryzen patch:

https://hardforum.com/threads/dota-2-update-increases-performance-for-ryzen.1927957/

The OP claims the patch improved performance by 23% on RyZen. The i7 7700k was a huge 50--60% ahead the R7 1800x before the patch

http://www.phoronix.com/scan.php?page=article&item=nvidia-1080ti-ryzen&num=2

But the problem is solved... ana 23% improvement is no small gain... So patches can an will make a big difference it seem's... Doesn't really matter if it catchs kaby lake... As the difference is impercievable anyway while playing that is...

The problem is not solved. What they did was to take the worst game and patch it. No one would expect 23% improvements for any patched game and no one would expect all games to be patched.

Glass half empty, there. And even if you didn't expect it, it just happened anyway.

Just accept that fact: games do get a big boost from patching for Zen. I do agree not all will see a massive improvement like that, but there will be improvements.

Problem or not, it doesn't matter. If they can patch games (or the OS scheduler) to make better use of Zen, you will see outliers like DOTA2 and then some games that might not even show a boost.

Cheers!

 

[salgado18]

Spoiler

Dota 2 before and after Ryzen patch:

https://hardforum.com/threads/dota-2-update-increases-performance-for-ryzen.1927957/

The OP claims the patch improved performance by 23% on RyZen. The i7 7700k was a huge 50--60% ahead the R7 1800x before the patch

http://www.phoronix.com/scan.php?page=article&item=nvidia-1080ti-ryzen&num=2

But the problem is solved... ana 23% improvement is no small gain... So patches can an will make a big difference it seem's... Doesn't really matter if it catchs kaby lake... As the difference is impercievable anyway while playing that is...

The problem is not solved. What they did was to take the worst game and patch it. No one would expect 23% improvements for any patched game and no one would expect all games to be patched.

Glass half empty, there. And even if you didn't expect it, it just happened anyway.

Just accept that fact: games do get a big boost from patching for Zen. I do agree not all will see a massive improvement like that, but there will be improvements.

Problem or not, it doesn't matter. If they can patch games (or the OS scheduler) to make better use of Zen, you will see outliers like DOTA2 and then some games that might not even show a boost.

Cheers!

Which is why talking averages is not good in this case. Many games will see no patch at all, and will pull the average down. A 5% improvement overall could mean games with problems reach 10-20% improvement or more, which is great.

 

[juanrga]

Hey all, I just ran my first bench test upon my first start with my Ryzen. I am using my stock AMD cooler from my old Phenom II x6 1035T (I have a Corsair H100i v2 closed loop water cooler, but the AM4 bracket isn't out yet). Here is my first test score on Single Thread and Multi Thread: ST= 2165 MT= 18720 on the CPU-Z test.

Note that CPU-Z is a toy bench; It doesn't represent realistic performance. Also CPU-Z has a bug that gives RyZen more performance than it really has.

 

[juanrga]

https://arstechnica.co.uk/gadgets/2017/03/amd-ryzen-gaming-performance-analysis/

A very good resume of the hype and the facts. It destroys most of incorrect claims have been repeated in forums and even some tech sites. I agree with most, except with a pair of claims, like some final guesses about the future of gaming. They look to me like a repetition of what was promised five years back.

 

[Redneck5439]

Dota 2 before and after Ryzen patch:

https://hardforum.com/threads/dota-2-update-increases-performance-for-ryzen.1927957/

The OP claims the patch improved performance by 23% on RyZen. The i7 7700k was a huge 50--60% ahead the R7 1800x before the patch

http://www.phoronix.com/scan.php?page=article&item=nvidia-1080ti-ryzen&num=2

But the problem is solved... ana 23% improvement is no small gain... So patches can an will make a big difference it seem's... Doesn't really matter if it catchs kaby lake... As the difference is impercievable anyway while playing that is...

The problem is not solved. What they did was to take the worst game and patch it. No one would expect 23% improvements for any patched game and no one would expect all games to be patched.

Glass half empty, there. And even if you didn't expect it, it just happened anyway.

Just accept that fact: games do get a big boost from patching for Zen. I do agree not all will see a massive improvement like that, but there will be improvements.

Problem or not, it doesn't matter. If they can patch games (or the OS scheduler) to make better use of Zen, you will see outliers like DOTA2 and then some games that might not even show a boost.

Cheers!

Many existing or older games may never see such patches or optimizations, but going forward most if not all big title AAA games should have AMD optimizations. For about a decade the only processors that mattered for gaming were Intel so by default games were optimized to run on the best platform aka iCore. When AMD got the contract for console systems we saw a lot of "growing pains" with games running on AMD hardware but relatively few true optimizations as again most games were developed with an Intel first approach (as that is what they were use to and Intel was way ahead of Piledriver). With Ryzen being so close to Intel when optimized and the console systems (the gaming industry's bread and butter) obviously going to Ryzen next generation most AAA titles should be made from the ground up optimized for Ryzen. That's not saying that they also won't run great on Intel, but the performance gap that we see in outlier games should be greatly reduced going forward. I would also suspect many more games using the Vulkan API which in many respects is superior to DX12. This will allow future games made for Ryzen based console systems perform the best that they can, as a side effect we will also see Ryzen processors benchmarking much better in computer based gaming scenarios.

 

[Redneck5439]

https://arstechnica.co.uk/gadgets/2017/03/amd-ryzen-gaming-performance-analysis/

A very good resume of the hype and the facts. It destroys most of incorrect claims have been repeated in forums and even some tech sites. I agree with most, except with a pair of claims, like some final guesses about the future of gaming. They look to me like a repetition of what was promised five years back.

The biggest performance boost I've seen with Ryzen is when they can get the memory running at a fast enough speed 3000Mhz+. With most workstation based applications the boost is minimal, however for gaming it gives Ryzen a nice boost in performance. Now I'm not saying that the bios, motherboard and microcode updates that are finally allowing Ryzen to achieve such RAM speeds are a "magic bullet" and now Ryzen is outperforming Intel. That is not the case, Intel is still ahead, but with the RAM speed increased the gaming gap is greatly reduced to the point where it becomes more and more a non-existent performance gap as you would never notice it while gaming unless you enjoy having a Fraps overlay on all your games breaking any immersion one might enjoy while gaming. With Ryzen RAM speeds seem to be the key to gaming performance. It will be interesting to see how well Ryzen can do in game as bios updates mature and higher RAM speeds are fully supported and stable.

 

[-Fran-]

Which is why talking averages is not good in this case. Many games will see no patch at all, and will pull the average down. A 5% improvement overall could mean games with problems reach 10-20% improvement or more, which is great.

You still need averages, but they do need proper context; that is a must.

A game behaving as an outlier is a symptom of something else usually. Like bad code or a lot of "upgrade" at once; this is in the context of "ponying up" for Zen.

https://arstechnica.co.uk/gadgets/2017/03/amd-ryzen-gaming-performance-analysis/

A very good resume of the hype and the facts. It destroys most of incorrect claims have been repeated in forums and even some tech sites. I agree with most, except with a pair of claims, like some final guesses about the future of gaming. They look to me like a repetition of what was promised five years back.

Good read. Thanks for the link.

Many existing or older games may never see such patches or optimizations, but going forward most if not all big title AAA games should have AMD optimizations. For about a decade the only processors that mattered for gaming were Intel so by default games were optimized to run on the best platform aka iCore. When AMD got the contract for console systems we saw a lot of "growing pains" with games running on AMD hardware but relatively few true optimizations as again most games were developed with an Intel first approach (as that is what they were use to and Intel was way ahead of Piledriver). With Ryzen being so close to Intel when optimized and the console systems (the gaming industry's bread and butter) obviously going to Ryzen next generation most AAA titles should be made from the ground up optimized for Ryzen. That's not saying that they also won't run great on Intel, but the performance gap that we see in outlier games should be greatly reduced going forward. I would also suspect many more games using the Vulkan API which in many respects is superior to DX12. This will allow future games made for Ryzen based console systems perform the best that they can, as a side effect we will also see Ryzen processors benchmarking much better in computer based gaming scenarios.

I don't disagree, but I still see games being patched; popular ones. Hell, even Starcraft 1 will get a patch in a few months time! I don't think they'll add support for Zen, obviously, but specific CPU optimizations are an exception rather than the rule and that is where I agree.

I've been checking in Steam all patch notes for the big games I play if they mention anything about Zen, and some mention stuff that might point to support for Zen, but I haven't been able to confirm. Some others might not even put it in a patch note.

Going forward is just the same Pandora's Box we're all accustomed to, I guess. I do hope Vulkan get wider adoption, but so far it's been slow for big AAA games at best. All of them are too damn tied to DX in one way or another. Those big executives need to pay their BMWs and all! Hahaha. Tinfoil hat not-withstanding, most big publishers are not going to port their biggest engines to Vulkan so easily if they already have them working in DX9c+.

Cheers!

 

[8350rocks]

Supposedly, the new microcode updates allow even faster RAM. Hopefully they unlock the sub timings on the boards so you can run tighter timings.

 

[jdwii]

Jayztwocents and gamernexus as well as Salazar Studio are all objective to Ryzen

https://www.youtube.com/watch?v=L4K7eIEAJx0&t=0s