There is only a single Zeppelin die variant. It features 2 CCXs, with four cores, 2MB L2 and 8MB of L3 in each.
All Ryzen models which have < 8C are harvested (downcored).
That is what I thought. So it should be quite possible to have a 8C chip that gets cut into 6C chips much like the x3 720 was cut down from 4C correct?
If so, interesting...
I'm in general impressed with just how good AMD's control of information has been, looks like they took the lessons from the Eyefinity situation to heart.
AMD currently validates Ryzen 6C / 12T @ 3.3 GHz base but NOT 8C / 8T. The goal is clearly to recycle the failed 8C
That's what makes me uneasy. This is very similar to Bulldozer when AMD was making bold claims but as we neared the launch they went tightlip as they realized BD sucked.
Hope its not the same here.
Except AMD didn't drum up anything with Ryzen. They've been incredibly tame from start to finish.That's what makes me uneasy. This is very similar to Bulldozer when AMD was making bold claims but as we neared the launch they went tightlip as they realized BD sucked.
Hope its not the same here.
There is only a single Zeppelin die variant. It features 2 CCXs, with four cores, 2MB L2 and 8MB of L3 in each.
All Ryzen models which have < 8C are harvested (downcored).
That is what I thought. So it should be quite possible to have a 8C chip that gets cut into 6C chips much like the x3 720 was cut down from 4C correct?
If so, interesting...
6c/6t would be a really interesting chip, if priced competitively and performance is there, to go after the i5, no? But if such a unicorn exists, I imagine it will show up at end of year?
Rumor has it that amd is last minute scrabbling to come 10% close of the i7 6900k.
Don't the benchmarks already show them there if not ahead? The last I knew, AMD was only 10% behind Kabylake IPC not Broadwell-E where they are at least tied.
I thought they were ~5% of the 6900K in that handbrake/blender...but that is only 2 benchmarks. AFAIK, no one knows anything (public) about overall perf.
I thought they were ~5% of the 6900K in that handbrake/blender...but that is only 2 benchmarks. AFAIK, no one knows anything (public) about overall perf.
But that was a lower clocked ES right? If the rumors are true, the higher SKU at least should be clocked to 3.6... I mean even CES samples had 3.6 base.
This. We just have to find which workloads are representative of handbrake/blender demos, and which are not.There are bound to be workloads other than what AMD has demoed where Zen is less capable vs Intel. Sounds like they want to charge $450+ for top end and are trying to get as close as possible to a 4GHz advertised boost to help soften the blow for potential customers hoping they would stick to thinner margins.
Crossing my fingers there is a sweet spot SR5 SKU at launch. I don't want to spend more than $300 on CPU but also want to play with a 6-8 core Zen chip preferably with ECC supporting motherboard.
This. We just have to find which workloads are representative of handbrake/blender demos, and which are not.
Vanilla blender rendering isnt too common, i would rather had had cycles rendering comparisons if we stick to blender as the 3d modeler.
Lets just wait and see how this pans out.
Sent from my XT1040 using Tapatalk
two benchmarks is not nearly enough to judge that.Don't the benchmarks already show them there if not ahead? The last I knew, AMD was only 10% behind Kabylake IPC not Broadwell-E where they are at least tied.
two benchmarks is not nearly enough to judge that.
Skylake/Kabylake IPC is only 3-4% higher than Broadwell. On the benchmarks of new horizon, Zen is moreless on par with a 3.4GHz skylake and so moreless with same IPC (slightly more in handbrake).Don't the benchmarks already show them there if not ahead? The last I knew, AMD was only 10% behind Kabylake IPC not Broadwell-E where they are at least tied.
English summary:
We have confirmed from our sources that AMD’s upcoming Ryzen processor can be technically configured as 6-core model. It is possible to disable each CPU core separately together with dedicated L2 cache from CCX without affecting shared L3 cache. Possible L3 cache configurations are 1/1 (8 MB), 1/2 (4 MB) or completely disabled. Basic rule is that both CCXs should have similar CPU core and L3 configuration for example both should have 3 cores and full 8 MB L3 cache enabled (6C & 16 MB L3).
We believe AMD will launch 6-core model of Ryzen CPU. If they don’t, it will be due other reasons than technical limitation of CCX.
Well were there multiple leaked benchmarks that showed Bulldozer in a positive light back then? Specifically compared to $1000+ Intel HEDT parts?
True. But I believe the two benchmarks AMD showed is where zen had the best performance, which is probably why they showed them and not others.
CPC benchmark suggested that Zen was about 8pc behind BWE clock for clock. Which is impressive coming from BD, but none of the them showed single threaded IPC. We don't know if AMD has better SMT or not. but if they do have a better SMT, it means single thread performance is more than 15 pc behind SL/KL at the very least. While its not bad, but it means to undo Intel at the very top, they will need to clock their 4c8t CPU more than 4.5 GHz at the very least. If IPC is 20pc behind, than the clock has to be 4.8ghz. And soon it becomes unlikely that a zen core can clock at 4.8ghz out of the box.
AMD has claimed 40+pc IPC in single thread over excavator. PD/XV is roughly 75pc behind SL on average.
Granted AMD don't to take the performance crown to have a winning product, but it will probably be helpful to win some serious market share. Please consider that most people will just look at the benchmark that will show Intel is 10 pc ahead and so get Intel instead as most likely they don't over clock.
Granted AMD don't need to take the performance crown to have a winning product, but it will probably be helpful to win some serious market share. Please consider that most people will just look at the benchmark that will show Intel is 10 pc ahead and so get Intel instead as most of them don't overclock.
AMD currently validates Ryzen 6C / 12T @ 3.3 GHz base but NOT 8C / 8T. The goal is clearly to recycle the failed 8C
Hmm so apparently Canard is strongly reputing the 8c/8t + no 6c/12t rumor... Interesting... More mixing of the pot