AMD Zen “RYZEN” CPUs Detailed – 8 Cores, 3.4Ghz+ & Auto Overclocking With “XFR”

[beginner99]

Those TURBO buttons, when pressed, turned down the performance of your computer. lol
https://en.wikipedia.org/wiki/Turbo_button

I really needed that as a kid on the 286 CPU as playing pakman at 8 or 12 Mhz was impossible

On-Topic:

Looks promising but too many unknowns. Hopefully AMD did not cheat by slowing down the intel CPU (Clocks, or using a build without certain instructions) or OCing their own CPU. Many things the can cheat with.

Then again a perfect example for high thread count. Would be more interested also in single-threaded or "low-threaded" stuff like games. The SenseMI features could theoretically help here if they really boost high when only 4 cores are used. But how do you limit this?

I mean Polaris supposedly has these cool power controlling features too yet many of the cards shipped with much too high voltage so that you could just save a ton of power by lowering voltage and losing nothing. So there that did not work very well.

Ultimately it comes down to price. I won't spend $1000 on a CPU. Period. Probably depends how many of these AMD wants to sell or how many orders they get from Enterprise segment. Even if top SKU is $600, I don't see them selling well (compared to quad-core i5/i7) as it's just too much or limited returns.

 

[cytg111]

Is price/performance a linear function?

Buy one 8350 for x money get y performance. Buy two 8350 for 2x money and get 2y performance......... Im going with a yes. (sorry).

 

[SpaceBeer]

So when you combine 2 x FX-8350 will you get 16 cores @ 4GHz, or 8 cores @ 8 GHz :d

 

[coercitiv]

Buy one 8350 for x money get y performance. Buy two 8350 for 2x money and get 2y performance......... Im going with a yes. (sorry).

Well, enjoy your two 8350 systems! Meanwhile, people looking for 2x 8350 performance in a single socket might have to contend with the following:

• buy a 16+ thread CPU, which may offer 2x performance in highly multi threaded scenarios. Will cost more than 2x 8350.
  
• wait for a 8 threaded CPU 2x faster than 8350, maybe even foot the bill for the R&D necessary for to get 2x ST performance. Might be a bit more expensive than 2x8350 though. A few billion more.

 

[cytg111]

Well, enjoy your two 8350 systems!

- psst, was joking.

 

[superstition]

Well, enjoy your two 8350 systems! Meanwhile, people looking for 2x 8350 performance in a single socket might have to contend with the following:

• buy a 16+ thread CPU, which may offer 2x performance in highly multi threaded scenarios. Will cost more than 2x 8350.
  
• wait for a 8 threaded CPU 2x faster than 8350, maybe even foot the bill for the R&D necessary for to get 2x ST performance. Might be a bit more expensive than 2x8350 though. A few billion more.

Actually, I've wondered why no one released a dual FX board for the mainstream since it has been around forever, with AMD releasing the E chips to be able to run in DP mode.

Two $90 8320Es at Micro Center plus a reasonably-priced dual socket board could have interested some folks. Pair it with a decent VRM system and both could run at around 4.3 GHz. Of course, cooling would be a little trickier. A dual-head AIO would work but wouldn't benefit from economy of scale. The FX isn't good for single-threaded performance but if clocks are kept within the process' efficiency range the power consumption isn't terrible for workstation use.

 

[itsmydamnation]

i think the market is to small to bother, it would have been awesome for someone like me ( cheap home vm server) but thats about it.

 

[superstition]

i think the market is to small to bother, it would have been awesome for someone like me ( cheap home vm server) but thats about it.

Well, clearly for servers it would have been optimal since a stock cooler would work. However, it would also have worked for rendering and encoding. The 8370E stock cooler would have been a better choice, though, since the fan included with the 8320E is frankly terrible.

 

[itsmydamnation]

Well, clearly for servers it would have been optimal since a stock cooler would work. However, it would also have worked for rendering and encoding. The 8370E stock cooler would have been a better choice, though, since the fan included with the 8320E is frankly terrible.

i think the last stock cooler i used was on a barton XP 2500+ ....lol

 

[superstition]

i think the last stock cooler i used was on a barton XP 2500+ ....lol

At stock the 8370E's cooler is adequate. The fan is much thicker.

 

[nathanddrews]

Hey, can we discuss the demo showcasing Dota 2 broadcasting? If I understood correctly, the Ryzen alone (no dGPU) was playing Dota 2 at 1080p ("max" or "ultra" was mentioned) while broadcasting 1080p on Twitch without dropping frames. There were really no other details given that I recall, other than the 6700K was substantially overclocked and struggle immensely. What does that tell us about Ryzen's GPU? I mean, we already know Intel IGP can't draw itself out of a box in less than 5 frames.

EDIT: 6700K, not 6900K.

 

[itsmydamnation]

Ryzen doesn't have an iGPU

 

[sandorski]

Hey, can we discuss the demo showcasing Dota 2 broadcasting? If I understood correctly, the Ryzen alone (no dGPU) was playing Dota 2 at 1080p ("max" or "ultra" was mentioned) while broadcasting 1080p on Twitch without dropping frames. There were really no other details given that I recall, other than the 6900K was substantially overclocked and struggle immensely. What does that tell us about Ryzen's GPU? I mean, we already know Intel IGP can't draw itself out of a box in less than 5 frames.

Wasn't it a 6700k dropping frames? They had 2 Intel systems, the faster one was performing about the same as Ryzen, the other one was dropping frames.

 

[sandorski]

Ryzen is looking good, real good. It will be interesting to see how it performs in Reviews without the auto-overclock turned off.

 

[Genx87]

Those TURBO buttons, when pressed, turned down the performance of your computer. lol



https://en.wikipedia.org/wiki/Turbo_button

Funny! Although I usually pushed the button until it gave better performance not looking at what the LED was saying.

 

[nathanddrews]

Wasn't it a 6700k dropping frames? They had 2 Intel systems, the faster one was performing about the same as Ryzen, the other one was dropping frames.

Yes, thanks! 6700K OC to 4.5GHz. I fixed my post.

Dota 2 7.0
1080p "full blown" settings
1080p 8500 "kay-bits" streaming

Ryzen (presumably stock) playing/streaming perfectly
6900K (presumably stock) playing/streaming perfectly
6700K (OC to 4.5GHz) playing fine, but streaming is a stuttering mess

 

[Blue_Max]

Hopefully AMD did not cheat by slowing down the intel CPU (Clocks, or using a build without certain instructions) or OCing their own CPU. Many things the can cheat with.

The speaker made a specific point of telling everyone in the live event that both chips were running at absolute STOCK configurations. Time will tell if she was honest about it.

 

[TheELF]

The speaker made a specific point of telling everyone in the live event that both chips were running at absolute STOCK configurations. Time will tell if she was honest about it.

Yes and choice of mobo very much influences this,they made a very specific point in not telling us the mobo model.

 

[antihelten]

The speaker made a specific point of telling everyone in the live event that both chips were running at absolute STOCK configurations. Time will tell if she was honest about it.

Technically speaking the Ryzen chip wasn't running at stock, since it had boost disabled and was thus only running at the base 3.4 GHz. Running at stock with boost enabled it should be even faster (by how much remains to be seen obviously).

 

[jelome1989]

Yes, thanks! 6700K OC to 4.5GHz. I fixed my post.

Dota 2 7.0
1080p "full blown" settings
1080p 8500 "kay-bits" streaming

Ryzen (presumably stock) playing/streaming perfectly
6900K (presumably stock) playing/streaming perfectly
6700K (OC to 4.5GHz) playing fine, but streaming is a stuttering mess

Yeah about that, is that how it really works? The 6700k can't handle streaming and gaming at 1080p at the same time? Seems hard to believe

 

[hojnikb]

It could, if you enable quick sync.

 

[[DHT]Osiris]

I didn't watch the stream. They compared twitch streaming on Dota2 @1080p on a skylake vs ryzen? What was the GPU? How taxed were the CPUs during this? Were they using comparable motherboards/memory types? Did they actually bring up anything at all about this?

 

[LTC8K6]

Technically speaking the Ryzen chip wasn't running at stock, since it had boost disabled and was thus only running at the base 3.4 GHz. Running at stock with boost enabled it should be even faster (by how much remains to be seen obviously).

It's the most interesting feature, and they couldn't demo it?

Why not run the bench with and without it? They could do it without revealing clock speeds if that's what they want kept secret.

 

[TheELF]

It could, if you enable quick sync.

Even without it,dota is not that demanding on the CPU,it all depends on how you configure your streaming software, if it runs at lower priority then your game then the recording will skip frames no matter how much idle CPU you have because task manager will think that that's what you want.

 

[TheELF]

It's the most interesting feature, and they couldn't demo it?

Why not run the bench with and without it? They could do it without revealing clock speeds if that's what they want kept secret.

Could be that within that TDP that clock is the max for all threads so turbo would not add anything,so it's more dramatic to show the test without turbo and let peoples imaginations run wild.