Official AMD Ryzen Benchmarks, Reviews, Prices, and Discussion

[lolfail9001]

The entire point is that giving the faster memory tighter timings without mentioning it makes the comparison incredibly deceptive, at best. Especially since it's generally easier to tighten timings when you're running memory at a lower clock speed.

Like most of memory comparisons, to be honest.

 

[coercitiv]

The entire point is that giving the faster memory tighter timings without mentioning it makes the comparison incredibly deceptive, at best.

So we agree

 

[Valantar]

Based on its weight, I'd say no. Too much dead space in that design. I would like an enormous head-sized, high-fin-density rad with um, noisy fans? That thing is designed to be silent, and it actually weighs less than a D15. Ryzen is going to put out 200W+ for some users, so . . .

Yeah, I know, I was just showing you the biggest heatsinks I've ever seen Wonder what the result would be if they stuck some fins on it, and added a fan mount or three. Those mini heatpipes are interesting, though. wonder how effective they are, given the size and not having wicks.

 

[PotatoWithEarsOnSide]

From 2133C15 to 2400C16 to 2666C16 to whatever to 3200C14.

Got any issues?

Kind of, yeah.
Comparing CL16 2133MHz with CL14 3200MHz doesn't isolate bandwidth or timings; its obviously going to have better performance.
If the 3200MHz is run at CL14, ALL of the other frequencies should have been run at CL14 as a starting point, but also at tighter timings too in order to check whether it is latency or bandwidth causing increased performance.

 

[VDA-RTR]

Ryzen will scale well with memory speed

This is a fake made from this image.

 

[lolfail9001]

This is a fake made from this image.

Ddddddde-BUNKED.

Granted, that chart still makes no sense.

 

[Valantar]

http://imgur.com/a/Xl4gN

MSI X370 motherboard manual slides

Comparing that to the Z270 Xpower Gaming Titanium gives some interesting perspective on Ryzen's OC potential.

This is all pure speculation, of course. But:
-We know the average 7700K might reach 5GHz. But most likely the "6" step is as far as it will go with an automatic OC. The "4" step should work for most chips. The same goes for the 6700K - more than 4.7GHz is a stretch, while 4.6 isn't uncommon.
-Given that MSI is trying to make things equal across their boards, that should give stable auto OCs in the ~4.2 GHz range. And I take this to mean for all cores, of course.

I'd be more than happy with that.

 

[Valantar]

Yeah.

But from a legal aspect if it's only a rebate it's fine.

It's not quite as cut and dried as that. Anti-competitive rebate deals are still anti-competitive, and still illegal. Pricing competition out of the market is illegal, no matter if it's through rebate schemes or price hikes for non-compliant customers.

 

[tamz_msc]

Comparing that to the Z270 Xpower Gaming Titanium gives some interesting perspective on Ryzen's OC potential.

This is all pure speculation, of course. But:
-We know the average 7700K might reach 5GHz. But most likely the "6" step is as far as it will go with an automatic OC. The "4" step should work for most chips. The same goes for the 6700K - more than 4.7GHz is a stretch, while 4.6 isn't uncommon.
-Given that MSI is trying to make things equal across their boards, that should give stable auto OCs in the ~4.2 GHz range. And I take this to mean for all cores, of course.

I'd be more than happy with that.

That'd be the same as what most Broadwell-E chips can do with a top of the line AIOs or a top air cooler like the Dark Rock Pro 3/NH-D15. It's all down to the voltages then.

 

[inf64]

Ddddddde-BUNKED.

Granted, that chart still makes no sense.

It would be hilarious if Ryzen actually gets that kind of performance in games(~top i7 KL) and scales like that with faster DDR4.

 

[lolfail9001]

It would be hilarious if Ryzen actually gets that kind of performance in games(~top i7 KL) and scales like that with faster DDR4.

Choose timings appropriately and it should.

 

[bjt2]

If you undervolt + underclock then you change the functional TDP of the chip. I can drop the power usage of my A10-7870k into the floor, but it still doesn't mean I can get one in a Dell laptop.



Yeah you'll get some boutique brands doing it. With extra-large batteries and such.

Consider that Zen is a SoC. And Sata, M.2, USB ports provided by the CPU are enough for a tablet/laptop.
So a Zen powered laptob should only have SSD, RAM and SoC, plus ancillary circuits...
With 15W TDP and just few chips, you will have the inner ambient very cool: STAPM will kick in and i think that with a quad core SMT APU can get 3GHz+ on 4 CPU and 1.2GHz+ on 512SPs...
If you have an HBM2 enabled APU, you can avoid RAM altogether (if 8GB oare enough) and save further TDP.
When Raven Ridge will come, we will have very powerful notebooks...

 

[lolfail9001]

Well, courtesy of everyone's favorite Spaniard, not sure if was posted yet.

http://ranker.sisoftware.net/show_r...e6d5ecdfecdbeec8ba87b791f491ac9cbac9f4c4&l=en
http://ranker.sisoftware.net/show_r...e4dceadceddae3c5b78aba9cf99ca191b7c4f9c9&l=en

Memory latency issues are real.

 

[Valantar]

Consider that Zen is a SoC. And Sata, M.2, USB ports provided by the CPU are enough for a tablet/laptop.
So a Zen powered laptob should only have SSD, RAM and SoC, plus ancillary circuits...
With 15W TDP and just few chips, you will have the inner ambient very cool: STAPM will kick in and i think that with a quad core SMT APU can get 3GHz+ on 4 CPU and 1.2GHz+ on 512SPs...
If you have an HBM2 enabled APU, you can avoid RAM altogether (if 8GB oare enough) and save further TDP.
When Raven Ridge will come, we will have very powerful notebooks...

Are you saying we'll have 15W 4c8t 3GHz APUs with 512SPs at 1.2GHz? At full load? Not power throttling? Don't be daft. At 45W that might be possible, but the GPU will throttle no matter what. Remember, the 896SP 460 is 75W alone.

 

[bjt2]

If you can keep the OS scheduler from putting threads onto them...

In Haiku OS that's easy to do live, Windows and Linux both make it a little more... troublesome... if not impossible to select exactly which core(s) to 'disable.'

The goal of that OC utility is to squeeze the maximum performance from the CPU. Just set each core to max supported frequency. Have a core shut off, let you gain little in frequency on the others and lose a lot in throughput (especially those tiny system services).
Don't zero the clock of bad cores. Just setup a daemon that each few seconds set affinity mask of all processes excluding slow cores. This process will fail on system processes (e.g. svchost and others), so your critical user processes will get always only the faster cores and the system and background tasks will use also the slow cores...

 

[R0H1T]

Are you saying we'll have 15W 4c8t 3GHz APUs with 512SPs at 1.2GHz? At full load? Not power throttling? Don't be daft. At 45W that might be possible, but the GPU will throttle no matter what. Remember, the 896SP 460 is 75W alone.

And the full P11 is 1024 SP @75W TDP, we also have the highly binned P11 for pros with a 50 W TDP ~
http://www.anandtech.com/show/10821/now-shipping-amd-radeon-pro-wx-series

I'm not saying that a 15W APU will throttle (or not) but you're basing your assumptions on a highly leaky, locked GPU.

 

[bjt2]

Are you saying we'll have 15W 4c8t 3GHz APUs with 512SPs at 1.2GHz? At full load? Not power throttling? Don't be daft. At 45W that might be possible, but the GPU will throttle no matter what. Remember, the 896SP 460 is 75W alone.

I calculated that many weeks ago and maybe i remember bad, but 15W excavator has 2.7GHz base clock for 4 core (2m) and 1100Mhz of (max?) frequency for the 512SPs. Given that at low power/Vcore the process allow 65% power saving or +80% clock, I calculated that if the SPs will remain 512 a +100MHz will give anyway some power to the CPU. 2.7GHz Zen 4C8T will draw just a little less than 2.7GHz XV 2M4C due to power scaling.
Giving the increased power budget on the CPU, we can reach 3GHz.
Or alternatively, still 2.7GHz Zen cores, but with 1024SPs at 900MHz-1GHz max, but with higher area consumption...

EDIT: to be clear, on the CPU the clock is base clock. I don't know/remember turbo clocks on the BR APU. For the GPU part it's the MAXIMUM clock. Base clock, if i remember well, is 800MHz... On the 28nm BULK.

 

[bjt2]

And the full P11 is 1024 SP @75W TDP, we also have the highly binned P11 for pros with a 50 W TDP ~
http://www.anandtech.com/show/10821/now-shipping-amd-radeon-pro-wx-series

I'm not saying that a 15W APU will throttle (or not) but you're basing your assumptions on a highly leaky, locked GPU.

What frequency are those 1024SP@50/75W? because i was talking of 512SP at a base frequency of 900-1000MHz on the 14nmFF (versus 512SP@800MHz on the 28nm BULK)

 

[french toast]

Well, courtesy of everyone's favorite Spaniard, not sure if was posted yet.

http://ranker.sisoftware.net/show_r...e6d5ecdfecdbeec8ba87b791f491ac9cbac9f4c4&l=en
http://ranker.sisoftware.net/show_r...e4dceadceddae3c5b78aba9cf99ca191b7c4f9c9&l=en

Memory latency issues are real.

Looks like it, the real question is does this even matter? We will find out one or another soon.

 

[bjt2]

I found the 12/15W APU specs (i was off by something): FX 9800P 4C @2.7/3.6 and 512SP@758MHz base clock (unspecified max)

So for RR 15W APU we can expect 4C@2.7/3.6 plus 512SP at 900MHz+
If you give something to the CPU, 3GHz base can be feasible...

 

[GroundZero7]

Well, courtesy of everyone's favorite Spaniard, not sure if was posted yet.

http://ranker.sisoftware.net/show_r...e6d5ecdfecdbeec8ba87b791f491ac9cbac9f4c4&l=en
http://ranker.sisoftware.net/show_r...e4dceadceddae3c5b78aba9cf99ca191b7c4f9c9&l=en

Memory latency issues are real.

That's DDR4 3000 running at CL16. That memory is rated for CL15.

69.5ns total latency...That's right on point. 3200CL14 is ~45ns..which is right about where my DDR3 1600 7-7-7-22 is at.

http://www.techspot.com/images2/news/bigimage/2015/09/2015-09-16-image-7.jpg

 

[inf64]

Well, courtesy of everyone's favorite Spaniard, not sure if was posted yet.

http://ranker.sisoftware.net/show_r...e6d5ecdfecdbeec8ba87b791f491ac9cbac9f4c4&l=en
http://ranker.sisoftware.net/show_r...e4dceadceddae3c5b78aba9cf99ca191b7c4f9c9&l=en

Memory latency issues are real.

I doubt "latency" will matter much. We already have numbers from numerous different workloads, Ryzen is very competitive. It can only get better I guess if the latency is brought a notch down.

 

[lolfail9001]

Looks like it, the real question is does this even matter? We will find out one or another soon.

Stuff i work with is generally unfriendly to any cache structure and consists of a load of random accesses.

So, for my workloads it probably does. I just hope AMD was aware of it ahead of time, so that stuff is fixed proper by now (in Zen 2, that should soon tape out, shouldn't it?). After all, that's about the only glaring flaw with Ryzen right now. Rest are either imperfections of deliberate choices.

That's right on point. 3200CL14 is ~45ns

45ns is latency on Intel CPUs. Having 25ns on top of that? Please.

I doubt "latency" will matter much. We already have numbers from numerous different workloads, Ryzen is very competitive. It can only get better I guess if the latency is brought a notch down.

Yep, cache should carry it for now.

 

[lightmanek]

That's DDR4 3000 running at CL16. That memory is rated for CL15.

69.5ns total latency...That's right on point. 3200CL14 is ~45ns..which is right about where my DDR3 1600 7-7-7-22 is at.

http://www.techspot.com/images2/news/bigimage/2015/09/2015-09-16-image-7.jpg

Where do you get 3000MHz memory speed from if SiSoft is reporting 2133MHz CL16?

 

[GroundZero7]

Where do you get 3000MHz memory speed from if SiSoft is reporting 2133MHz CL16?

CMK16GX4M2B3000C15 is
https://www.newegg.com/Product/Product.aspx?Item=N82E16820233852