yea 1hz = 1hz.. back in the day amd marketed their procs with performance ratings instead oh Hz's ... not today..
yea 1hz = 1hz.. back in the day amd marketed their procs with performance ratings instead oh Hz's ... not today..Something I've been wondering: Will 1 Ryzen Ghz = 1 KL Ghz? Not asking to cast doubt on Ryzen, I'm just curious if anyone would know because I remembered back in the old days (Thunderbird era or so) AMD used to market that one of their mhz was better than one of Intel's. We see leaks of 3.6 base release and whatnot but is that equal to 3.6 in Intel numbers too?
If they can't harvest 4 out of 8 cores, they messed up big time... No way you create a design like this and them gimp your binning prospects in the process.
AMD actually harvested half of Polaris 11 that is tiny on it's own.
We already knew that without the need to link to that site
and this might not be final silicon?
i dont understand what you are saying?That s for AVX2 but if there s only 128bit loads , like AVX128, then it will be one 256b load and one 128b store, because it doesnt have the ressources to exe two 128 bit loads using two different instructions, we are talking of SIMD here, so we are back to the same number as Zen.
That price without Intel's AVX-2 performance?And for those clock AMD can even charge 600-800$ for the top end SKU...fine with me!.
AVX-2?I'm not sure people realize how fast this 3.6/4Ghz part will be . It will be brutal across the board and likely good deal cheaper than any 8C intel part.
It has in some instances. Chasing lower price points has caused board makers to lower the standards, rather than raise them, in some board revisions. The third 990FX Sabertooth board, for instance, has a VRM sink designed for a 40mm fan. The drawing I saw even showed that airflow is supposed to go backward, too — into the case. L-O-L.I swear to god, sometimes I think mobo development is going backwards, ffs.
Is the 4300 really very useful for comparison, since it's not a full Piledriver CPU? I think those charts should include an 8320E or 8370E.I could argue that there's no difference. In R11.5 FX-4300 is about 14% faster than i3-4130T. In R15 it is also 14% faster than i3-4130T. Conclusion, no bias towards Intel. But what we could get from these results is that cache size matters for R15. X4 845 has half the L2 per core compared to FX-4300. In R11.5 845 is about 6% faster than FX-4300, but in R15 it is exactly the opposite(6%). So half the L2 hurts X4 845 for about 12% in R15. Roughly the difference between your comparison between R15 and R11.5 results.
Umm, P11 has 16CUs, meaning 1280 shader units. The RX 460 has two of them disabled, so 14CUs and 1024 shaders.So the most cut down version is half of P11.No, they didn't. The P11 die has 1024 shaders; the most cut-down version we've seen is the Radeon Pro 450 with 640 shaders, and that's a Macbook Pro exclusive, so it was probably done at Apple's express request. The furthest cut-down version that has made it into an AIB product was 896 shaders (only 12.5% cut-down).
No matter how much I like SIMD (where's my preciooouus AVX-512? ), I don't think the impact of AVX2 is that important at the moment for the vast majority of people.AVX-2?
>Haswell can perform 256bit loads or store.
Well, sense with what resources they have.
But if AMD had more manpower I'm sure they could release more and sooner. I mean with the looses over the last decade it's amazing they have (hopefully) made such a comeback.
However, if they had spare cash I'm sure that once they had accurate projections of Zen's performance they would have hired lots of people for validation etc. Risky and might have been unpopular with shareholders (hiring before revenue), but if it would have knocked of a few months (or more for the lower end parts) it would probably be money well spent.
I think currently AMD are rather stretched for everything which was also noticeable with the Polaris launch where I'd guess the PS4 Pro work held up the launch. Would not be surprised if the Project Scorpio work is delaying Vega.
Obviously some things like tapeouts take time and simply throwing more people at problems doesn't always work, but in this case it probably would have made sense.
I think we are saying the same thing.As said only with AVX2 because the load is 256bit, but that s the same registers that are used for 128bit loads, in that latter case you cant use the 128 upper bits to load a second and separate 128bit load, because the 256bit load is attached to a single AVX2 instruction while two separate 128bit loads would be attached to two separate instructions that would require two separate registers.
AVX-2?
There is also asia specific 470D, which has 1796SPNo, they didn't. The P11 die has 1024 shaders; the most cut-down version we've seen is the Radeon Pro 450 with 640 shaders, and that's a Macbook Pro exclusive, so it was probably done at Apple's express request. The furthest cut-down version that has made it into an AIB product was 896 shaders (only 12.5% cut-down).
On the bigger P10, the shader count goes down from 2304 on the full-die RX 480 to 2048 on the cut-down RX 470. That's only about a 11% cut.
If similar yields appear on Zen (and they should be better, since it's now over 6 months after the Polaris launch), then there would be no reason for AMD to offer anything less than a 6-core version.
Guess you need to go back to preschool, because you math doesn't check out.Umm, P11 has 16CUs, meaning 1280 shader units. The RX 460 has two of them disabled, so 14CUs and 1024 shaders.So the most cut down version is half of P11.
Here it might help to split AVX2 benefits in new instructions, more efficient 3 operand mode for int SSE2 stuff and adding 256b to the latter.No matter how much I like SIMD (where's my preciooouus AVX-512? ), I don't think the impact of AVX2 is that important at the moment for the vast majority of people.
arg this bugs me so much (it important to get the semantics right dammit!)
Its not AVX workloads, its 256bit ops, any 256bit ops, meaning both avx and avx 2 can target 128bit vectors if they want. For example the extra instructions in avx/avx2 vs see, or because their console code is tuned to 3 operand 128bit avx etc. ( first thing that bugs me out of they way)
Also its not really the width of the units that's the limiting factor for Zen because it has more FP units then skylake. Its the load store bandwidth in and out of the core's. Zen has 256bits load and 128bits store vs 512/256 of >Haswell.
The point being AMD AVX and AVX2 performance is fine, there isn't some magically thing making AMD crap at those instruction sets( BD and PD had real 256bit instruction issues). Its that intel have an advantage on anything thats at 256bit operation. At the same time AMD/ZEN have an advantage on 128bit operations because they have more units.
If i was amd i wouldn't go chasing 256 or 512bit avx performance or SMT 4*, i would be using the massive die and power budget those things cost to increase clocks and IPC. If you look at AMD GPU's ( or NV) they are becoming much better at being CPU like. The more GPU's compute capacity becomes flexible and general the more a 512bit CPU becomes a jack of all trade master of none. If you have a master at both you can just eat them from both sides.
The General server base doesn't care about really wide vectors, thax to intels own segmentation neither does the consumer market.
*those rumors from Fottemberg aren't worth the bits in the database they are stored on unless AMD plan on basically copying a power9 style methodology to core design which is really a more unified version of CMT.
That just means f4 stepping has been completed not that it's the final stepping.Yes it is. Its releasing soon, they will not do a respin. The 6900k turbos at 3.7
Quote:
"The good folks over at Canard PC Hardware also revealed that the Ryzen F4 stepping has already been finalized and has a turbo clock of 4.0 GHz"
I think we are saying the same thing.
>haswell can do two loads a cycle upto 256b per load but a single load instruction might only be 32b,64b,128b etc
Zen can do two loads a cycle upto 128b per load but a a single load instruction might only be 32b,64b,,etc
Yet, as soon as Zen hits suddenly there could be a push to make it important. AVX-2 has been supported since Haswell on the Intel side, right?No matter how much I like SIMD (where's my preciooouus AVX-512? ), I don't think the impact of AVX2 is that important at the moment for the vast majority of people.