I highly doubt Intel could, server market look closely at perf/cost(price/power) so If AMD offer a better pref/cost product companies will buy them. Intel can have higher cost due to brand power and SW optimized for them but not by a high margin. Also in 2018 we will see Zen+ so AMD will have HW to compete. I don't expect the real answer to ryzen until 2019/20 where Intel will have time to develop and arq aim to fight zen instead of tweaking existing ones.
X denotes the XFR turbo https://www.techpowerup.com/230609/amd-ryzen-xfr-frequencies-revealed For regular overclocking it shouldn't make a difference, apart from binning of the SKUs
I thought the XFR turbo thingy was a feature that would auto-overclock your CPU to whatever your cooling could handle? Or at least I thought I read that somewhere.
I never trust auto-OCing anythings, never had much luck with them and always seem to do a lot better a lot faster by just adjusting everything manually. Here's hoping that will change.
I never trust auto-OCing anythings, never had much luck with them and always seem to do a lot better a lot faster by just adjusting everything manually. Here's hoping that will change.
Silent_Buddha
Legend
It's not so much the brand but proven stability, support, and performance over time. That all ties into the brand, but isn't exactly the same.
Basically the server market is generally risk adverse. That makes sense as gearing up for alternate architectures (ARM, for example) is expensive. There's a lot of legacy components (both software and hardware). IHV support contracts are negotiated sometimes for multiple years. Etc.
Zen being x86 will help. But similar to Opteron before it, it could take a year or more after introduction into the market before it gains significant traction. That's not a big deal, as initial supply wouldn't be able to meet overwhelming demand (more than a few percentage points of the market switching to AMD) anyway.
It's important for AMD to make a good showing there as the server market (HPC as well) are very high margin markets. And AMD really needs some high margin products to move in appreciable volume. Hence, why it's their primary focus for launch with consumer products derived from those.
Regards,
SB
thanks then the non x is the way to go for me , save that money for a better water cooler. Really sweet. Trying to tie some money together for an engagement ring this summer but then I can build myself a new pc lol
You don't have time for the PC anymore... especially if you do a kid immediately... then you don't have time to sleep either
QFT!
Also, no money. Not a problem though, because you don't have time to spend it.
Cheers
I bought a brand new Core 2 Quad + Radeon 5870 month before I met my future wife. It took a loooong time until the next update... Skylake 6700K. Nothing in between
D
Deleted member 13524
Guest
Wow, for an article titled "XFR frequencies revealed" that table gives away absolutely nothing. All it says is XFR models can give more than the Turbo values on XFR.
Perhaps there's no actual limit on the multiplier, and as long as voltages are stable and temperatures are below certain thresholds the mutliplier will just keep increasing.
So a phase change cooler together with a superb PSU and superb motherboard could have the CPU working at >5GHz with XFR.
Naah, I don't think so.
It's exactly that.
People will probably still get better results if they're willing to dedicate more heat + power + time + patience in their hands (at the very least because they will be willing to risk going over AMD's default voltage boundaries). XFR is for people who can afford more heat and power but don't have the rest.
Wow. Those are very nice clock rates for the Zen 8-core 95W TDP processor. 3.6 GHz base, 4.0 GHz turbo.
In comparison Core i7-6900K (8-core) base clock = 3.2 GHz and turbo = 3.7 GHz and TDP = 140W. AMD has 300-400 MHz clock rate advantage at 45W less TDP. This is awesome if true.
That 65W 8-core Zen (3.0 GHz base, 3.7 GHz turbo, 320$) is also interesting. Top of the line Xeon D 12-core (1300$) is 65W and 8-core (580$) is 45W, but both only have 2.1 GHz base clock and 2.7 GHz turbo. Against the 8-core model, Zen has 43% clock advantage, but also 44% higher TDP. Perf/watt should be very close. It would be a big deal if AMD could meet Intel perf/watt at 43% higher clock rate (significantly higher IPC). 12 core model has 50% more cores at same TDP as 8-core Zen. But as multi-core scaling is never perfect, I would assume 8-core Zen to be very close in performance. This of course all assumes that maximum load results in power consumption = TDP and all CPUs maintain base clocks (no throttling, no turbo).
Zen quad-cores on the other hand are underwhelming against the 4.2 GHz Kaby Lake (4.5 GHz turbo). But Kaby is 2 generations newer architecture than Broadwell.
Last edited:
You don't have time for the PC anymore... especially if you do a kid immediately... then you don't have time to sleep either
Sigh lol . I do a lot of home video editing on my rig so the hope would be that the upgrade would have me spend less time on the pc . Also yea kids will be tough but we are going to wait about 2 years so like 2020 for the kid. So I still have time for gaming !!!!
D
Deleted member 13524
Guest
To be fair, the 6900K has twice the memory channels and is based on the 2 generations-old Broadwell architecture, which like all "tick" transitions it doesn't behave all that well for higher clocks.
Skylake improved that a bit and so did Kaby Lake, which is why you get:
^^ Here they're indeed fighting the latest and greatest architecture from Intel.
I guess this lag in architectures between consumer and prosumer/HPC CPUs in Intel's portfolio is the main reason why AMD targeted the later first.
Different TDP, different binning.
Similar threads
- Locked
