This is actually how turbo works at default on my board. It sets the multiplier at 36x under any load which is the all core turbo multiplier of that CPU. It's called multi-core enhancement and all high-end boards have this and some set this as the default setting like mine does.On a second thought, What might happen is AMD will leave the multipliers "unlocked" up to a point where a given SKU is allowed to turbo. So a CPU with base/turbo of 3.0/3.4 GHz might have multiplier "unlocked" up to x34 in the BIOS/EFI.
So . . . are you allowed to tell us what is the VRM configuration on the C6H?
I cannot disclose any technical details.
However you can tell from the pictures in those slides that it is "sufficient"
Ok. And how much do you think it can sell a year?It's still too early to say but my guess would be that Snowy Owl would be at least $2500 to start, including the board.
Because that was the only way for those chips to compete. For AMD's sake I hope Zen is more competitive than FX.FX 8300, FX 8320, FX 8350 and FX 8370 all are unlocked. Same can happen with RYZEN.
On a second thought, What might happen is AMD will leave the multipliers "unlocked" up to a point where a given SKU is allowed to turbo. So a CPU with base/turbo of 3.0/3.4 GHz might have multiplier "unlocked" up to x34 in the BIOS/EFI.
FX 8300, FX 8320, FX 8350 and FX 8370 all are unlocked. Same can happen with RYZEN.
Because that was the only way for those chips to compete. For AMD's sake I hope Zen is more competitive than FX.
Marketing bullsheit.
Okay, so if we have two 8c/16t variants, and I have a top end board, what will be the reason I might buy one chip over the other, if both are fully unlocked?AMD has already stated that all Ryzen SKUs are fully unlocked.
While the CPUs themselves are unlocked, you will need a motherboard with a chipset variant (B350, X300 or X370) which has the overclocking feature unlocked.
No different to Intel's chipset segmentation, where only Z-series consumer chipsets support overclocking regardless of the installed CPU SKU (K).
On Ryzen base and boost states are completely isolated and the latter are completely invisible outside the SMU. So don't expect that you can trick the CPU to maintain the maximum boost speeds on boards which don't support overclocking.
Okay, so if we have two 8c/16t variants, and I have a top end board, what will be the reason I might buy one chip over the other, if both are fully unlocked?
It has to just be binning that would justify different pricing. One happens to auto overclock higher than the other with the same cooling.
But, AMD has no idea what cooling I might use, and therefore should not know what auto clock speeds I will get for my money. So how will they price them?
With Intel if I want to overclock a desktop system, I basically have three chip choices today. 7350K/7600K/7700K which is really only a single top chip, with 8 threads.
If I want to overclock a HEDT, I have 4 choices. 6800K/6850K/6900K/6950X with only one top chip, with 20 threads. Or if we eliminate the extreme edition, then one chip with 16 threads.
Intel does not offer two versions of the top chips. There's only one top chip with the most threads.
If Intel did offer such, they would have different base/turbo clocks and that would be the main reason to pick one. So that's Intel's version of binning, I guess. But that's how Intel would set the pricing.
Marketing bullsheit.
AMD currently validating Ryzen 6C/12T@3.3GHz base but no 8C/8T. The goal is clearly to recycle the failed 8C dies.
If that's achieved at 65W(as rumoured) TDP and has unlocked multiplier (as promised) then I don't see the problem for enthusiasts. Plus system integrators will love it because of low TDP.As long as validation only means a representative CPU that is checking for errors and not the final performance ceiling.
3.3G on a 6C12T wouldn't really cut it at the top end.
Okay, so if we have two 8c/16t variants, and I have a top end board, what will be the reason I might buy one chip over the other, if both are fully unlocked?
It has to just be binning that would justify different pricing. One happens to auto overclock higher than the other with the same cooling.
But, AMD has no idea what cooling I might use, and therefore should not know what auto clock speeds I will get for my money. So how will they price them?
With Intel if I want to overclock a desktop system, I basically have three chip choices today. 7350K/7600K/7700K which is really only a single top chip, with 8 threads.
If I want to overclock a HEDT, I have 4 choices. 6800K/6850K/6900K/6950X with only one top chip, with 20 threads. Or if we eliminate the extreme edition, then one chip with 16 threads.
Intel does not offer two versions of the top chips. There's only one top chip with the most threads.
If Intel did offer such, they would have different base/turbo clocks and that would be the main reason to pick one. So that's Intel's version of binning, I guess. But that's how Intel would set the pricing.
Nobody appears to know much as far as SKU's go. AMD managed to keep the lid on that quite well.Honestly I have no idea what the final lineup will be. Zeppelin is rather flexible, so there a plenty of potential configurations.
Honestly I have no idea what the final lineup will be. Zeppelin is rather flexible, so there a plenty of potential configurations.
Is it possible to have 6C/6T that is cut down from 8C/16T cpus?
Naturally.
The design dictates that both CCXs are enabled (three cores each), however besides that there are no limitations.
Which should actually just be binning alone in the case of RyZen, since the processor itself determines it's clocks.
There should not be any AMD locks or settings on the clocks. Each one should be determining it's own clocks and overclocks.
Unless AMD is releasing locked chips.