That power envelope wasn't gimped either.. I had no clue that sucker used 175Watts at peak performance . Given this, the conversations about 9900k have been even more absurd. If someone ever put an inefficient processor like this into a scaled data center they'd be suicided shortly after... And here I was thinking my 16 core threadripper could use less energy... Clown-tel running their 8 cores at this wattageI think that the main takeaway from the CB demo was that an 8c Ryzen 3xxx matched an 8c 9900K at ~4.7GHz. The implication is either the IPC is ahead, the clocks are ahead, or both are ahead but to a lesser extent.
The 9900K clearly wasn't gimped; it's CB score was comparable to those posted publicly by owners in the past.
Double FP performance across the board only applies to AVX2. Which CBR15 does not use.
My general assumption Is that the ES is running in the 4.0-4.3 GHz range, so we should not expect more than 300-400 more MHz out of a launch Zen2 product for top boost clocks. Feel free to prove otherwise, AMD!
My point was that it's a silly hotrod processor that is way beyond an efficient power envelope. 160W is about "normal".. Well, I'm glad I own a processor w/ double the I/O, memory bandwidth, and core count from a prior generation from AMD whose stock normal is only 180W....Tsk, nobody is putting a 9900k in a datacenter. 160W is about "normal" for a 9900k though unless you have a board that is configured to use that 210W PL2 value out-of-box (which some of them are). So 175W is not outside the realm of possibility.
That power usage on the Zen2 sample sure makes my 1800x look like crap, though.
I don't think we are in a position to guess what clock speeds the Zen 2 sample was running at.
4.0-4.3 is what I said. You mentioned ES at 4.6.
My point was that it's a silly hotrod processor that is way beyond an efficient power envelope. 160W is about "normal".. Well, I'm glad I own a processor w/ double the I/O, memory bandwidth, and core count from a prior generation from AMD whose stock normal is only 180W....
Pushing those kind of clocks on 14nm was beyond stupid and it shows in how absurd the power envelope is.
A 2700X @ 4.3GHz scores approx 1950 in CB15. Hypothetically, 4.5GHz would be enough to match the 2050 CB15 score.
Unless the IPC gains are far less than expected (people are expecting 10 - 15% right?) then this sample isn't running much higher than 4.0GHz
In power consumption, yes, unless you believe AMD will publish the better Intel score when the numbers are so neck to neck. Also, the enhancements AMD made to Ryzen 2 favors this kind of code, not to mention Cinebench is AMD's strongest suite in rendering benchmarks?What the hell? They showed a 75W CPU beating the best desktop CPU Intel has for offer that draws almost double the power when you exclude IO. And you find this disappointing?
Performance is EXACTLY asexpectedhoped for. That gap between 75W and 105-135W TDP will most definately be filled with extra performance by their top of the line products.
My own crude calculations have it as 4.2-4.3GHz Ryzen versus a 4.5-4.6GHz 9900K. I'm not sure how the 9900K would be capping out at 4.5-4.6GHz instead of going 4.7GHz ACT for the duration, as per stock settings, but the CB results do fit the slightly lower clocks.Double FP performance across the board only applies to AVX2. Which CBR15 does not use.
My general assumption Is that the ES is running in the 4.0-4.3 GHz range, so we should not expect more than 300-400 more MHz out of a launch Zen2 product for top boost clocks. Feel free to prove otherwise, AMD!
That power envelope wasn't gimped either.. I had no clue that sucker used 175Watts at peak performance . Given this, the conversations about 9900k have been even more absurd. If someone ever put an inefficient processor like this into a scaled data center they'd be suicided shortly after... And here I was thinking my 16 core threadripper could use less energy... Clown-tel running their 8 cores at this wattage
I think that the main takeaway from the CB demo was that an 8c Ryzen 3xxx matched an 8c 9900K at ~4.7GHz. The implication is either the IPC is ahead, the clocks are ahead, or both are ahead but to a lesser extent.
The 9900K clearly wasn't gimped; it's CB score was comparable to those posted publicly by owners in the past.
Thank you for the correction. My head almost exploded at the thought of it being otherwise.175W was the system power. The CPU uses about 125W. Still way more than the Zen2 ES but let's keep it sane...
To summarize my interpretation: The Ryzen 3 series uses the EPYC chiplets and a cut down version of the EPYC IO chiplet. The current 8 cores may end up being 8/12/16 depending on heat and power draw. The IPC is better, at least 10% more likely 15-17%. The clock speed seems to be in question, but most likely is higher than Zen+ 2000 series in the area of 4.7 maybe more in the final, but at least 4.4. The current power draw is almost half of Intels 9900k, but at what speed ? (125 vs 75, where 67.5 is half)
Sound about right ?
Higher memory speeds consume more power, so if they were using some (by Ryzen standards) ridiculously high memory speeds that would make the energy efficiency of the cores even higher.Another issue is: what memory speed did they use to run the Zen2 CBR15 bench? My 1800x defaults to DDR4-2133 and performace is terrible at those memory speeds. Jack it up to DDR4-3200 and it is much faster. The 2700x will default to 2933 if I am not mistaken. Assuming the demo was carried out using 1x DDR4, the default JEDEC spec is also going to be 2933. "Real" Zen2 samples @ launch may honor JEDEC specs for 1y DDR4 products which - if I recall correctly - should have standard speeds all the way up to DDR4-4266!
The IOC is confirmed to be a distinct die specific for desktop, Su said so in her keynote, and Ian Cutress in his AT article confirmed that the die is larger than a quarter of the Epyc IOC.To summarize my interpretation: The Ryzen 3 series uses the EPYC chiplets and a cut down version of the EPYC IO chiplet. The current 8 cores may end up being 8/12/16 depending on heat and power draw. The IPC is better, at least 10% more likely 15-17%. The clock speed seems to be in question, but most likely is higher than Zen+ 2000 series in the area of 4.7 maybe more in the final, but at least 4.4. The current power draw is almost half of Intels 9900k, but at what speed ? (125 vs 75, where 67.5 is half)
Sound about right ?
Underpromise. Overdeliver.I wonder why they didn't demo a higher core cpu off the bat? Perhaps they just wanted to compare the 8c to the 9900k and show off the improved IPC. Very interesting, mid 2019 is going to be great.
Or they took a cherry-picked, top binned ES 8C chiplet and used it for this preview.The other big takeaway was that 8c is in fact a fully functional chiplet, not 2 defectiv chiplets combined.
It may be that all X class SKUs are like this, and the non-X end up being defective chiplets. Reason being, fully functional 8c chiplet for 8c Ryzen actually suggests incredibly high yields.
I could only listen to about 10% about Dr Su CES presentation and it was about 7nm gaming and console (guessing vega) but writeup seems to confirm one chiplet for the 3rd gen Ryzen.To summarize my interpretation: The Ryzen 3 series uses the EPYC chiplets and a cut down version of the EPYC IO chiplet. The current 8 cores may end up being 8/12/16 depending on heat and power draw. The IPC is better, at least 10% more likely 15-17%. The clock speed seems to be in question, but most likely is higher than Zen+ 2000 series in the area of 4.7 maybe more in the final, but at least 4.4. The current power draw is almost half of Intels 9900k, but at what speed ? (125 vs 75, where 67.5 is half)
Sound about right ?
They might even hold off on launching anything above 8 cores initially. If supplies are tight, they might be able get more value out of using two 6-core dies to sell two CPUs instead of combining them into a single CPU.
now the I/O die is not a little too big for what it is?
At minimum should be 64 PCI-E lines, 2 DDR4, 4 sata, 4 USB 3.1, 2 IF links and thats about it.