The performance and efficiency improvement is >2x. Between that of Broadwell and Skylake.
Well maybe you expect Intel to continuously release "Core" architectures.I guess you will be disappointed with such expectations .
I guess you will be disappointed with such expectations .
But you'll let us know if you're going to manage to 2x performance and 2x efficiency at the same time, so we can go all in on Intel before everyone else does? (and then we can buy more Intel processors! :biggrin
You should expect a significant (emphasis not mine) increase in performance, in battery life, in power efficiency, all on this new product.
I'm excited-- In fact, I'm ecstatic on the health of Skylake.
Tejas/Nehalem (Netburst) would have used Enhanced Hyperthreading. Which has only been used in Poulson (Itanium).
.
What was so special about Titanium hyperthreading? Nehalem's reintroduced HT was already a big improvement over Netburst, and it has gotten even better since.
It seems that HT is giving us less gains in each generation as utilization of the units improve. HT is just a way to utilize unused units it's better to utilize them with just one thread because such an approach also improves ST performance. If they want to increase throughput very badly they can do 4 threads for a core just like power architecture does. Such an approach to HT is useful for servers but much less useful for desktops where ST performance is still the most important aspect of a CPU.
It seems that HT is giving us less gains in each generation as utilization of the units improve. HT is just a way to utilize unused units it's better to utilize them with just one thread because such an approach also improves ST performance. If they want to increase throughput very badly they can do 4 threads for a core just like power architecture does. Such an approach to HT is useful for servers but much less useful for desktops where ST performance is still the most important aspect of a CPU.
Well they added two execution ports with haswell. It is safe to assume they will do something similar with skylake. Perhaps more, if they dont dump a whole bunch more transistors into AVX. And I hope they dont, because they are dumping an awful lot of transistors into AVX, and its still really not being used. I assume intel realizes this too, and so they are hopefully giving that a rest for now and focusing more on general computing performance.
Another thing they could easily do, with so many execution ports, is add another thread to hyperthreading.
The latest Itanium Hyperthreading is and was Tejas/Nehalem (Netburst)'s Enhanced Hyperthreading.What was so special about Titanium hyperthreading?
If Skylake-S will be faster than Broadwell-K, then Broadwell-K is pretty pointless given that they will be released so close in time.
Sure, Broadwell-K can be overclocked. But the share of users actually overclocking their CPU should be very small. So for most people buying Broadwell-K will not make sense.
Previously, the K version was always released on the latest uArch and process node. So then it made sense to buy it even if you did not intend to overclock (it was the top end SKU regardless). But with Broadwell-K vs Skylake-S, this is no longer the case.
Does anyone have any opinion on this? If the logic is correct, doesn't it mean Skylake-S will have lower performance than Broadwell-K (even if not overclocked)?
It's because there won't be any Skylake-K CPUs at launch, or possibly at all. Calm down. You're overthinking it.
No, I mean that there won't be unlocked Skylake CPUs, based on Intel's road map.Well, if the slide mikk dug up is true, then there will be an LGA Skylake: http://forums.anandtech.com/showthread.php?t=2406498
Also, a 95W top bin either means useful gains in throughput (clocks and/or IPC) or AVX512 inclusion (which will likely bump up power consumption when running highly vectorized FP loads).
Well, that's a result of roadmaps not extending into 2016.No, I mean that there won't be unlocked Skylake CPUs, based on Intel's road map.
... And Skylake and Broadwell-K both release in 2015. My point stands. Broadwell-K will be the unlocked option for next year, and in 2016 it'll be replaced by either Skylake-K or, more likely, Cannonlake-K.Well, that's a result of roadmaps not extending into 2016.
Well, that's a result of roadmaps not extending into 2016.
Actually HT gains improved with Haswell thanks to the wider core.
Actually HT gains improved with Haswell thanks to the wider core.
That's what I deducted but I didn't see actual tests. I wish I had done some HT scaling tests when I had my 2600K system assembled just to be sure.Haswell core is wider but at the same time more efficient in regards of branch mispredicts and thread stalling. Haswell HT performance in MT is a regression than IVY/SB HT performance.
I have seen tests, and for example in the rendering engine I use (vray), HT has gotten progressively worse each generation as each core has become more efficient and wider.
Another way to check this is by looking into the first atom's HT implementations. In order atom was so terrible that HT on them felt like a breath of fresh air in terms of performance, night and day.