Your simulations start out as throughput-oriented but become sensitive to latency once it reaches steady state temperatures as time progresses? Also why should the CPU throttle after steady state - do you turn down cooling or what?Hyperthreading, when it works, works quite well but for a limited time. Then after hours of heavy loads hyperthreading often gets worse and worse (not usually tested in benchmarks that have a very short period in which to rush out the review). In my workstation experiences (where my simulations may run for weeks at a time with the CPU pegged at 100% utilization), hyperthreading just peters out as the chip thermally throttles.
It might do fine for a quick boost here and there. But for actual heavy loads, it isn't quite what people want.
Just to make sure we're on the same page, my sarcastic reply was in the context of expecting MT turbo to go up, not 1-4 threads turbo. The original claim was for chip throughput. If this was clear from the start, explanation is bellow.Not sure if sarcasm, but if it is, why do you think this isn't plausible?
They are being conservative to protect the value of the more expensive 14C/16C/18C models. There's a 12C Xeon Gold at 165W TDP with 3.2 GHz base / 3.9 GHz all-core Turbo. I still expect Core i9-7920X to have an all-core Turbo equal/above 3.4 GHz, so it will no doubt be faster than Core i9-7900X in MT tasks.
What about 8700k pricing?Coffee Lake tidbit of the day: It's a lot harder to pick a favorite from the new lineup than it was in the Kaby Lake generation, where Core i7-7700K is the clear choice for most people. Core i7-8700K is an achievement in its own right and a (very) worthy successor, but Core i5's performance is closer to the fastest SKU than it was before (based on Turbo).
Do you see any reason that Intel will vary much from the mid $300s price point that they have always targeted for this chip placement?What about 8700k pricing?
We saw that base clock a while ago.
Have to say i am surprised, did not expect it to be under 3GHz. I anticipated very close clocks to 7900x, maybe 100-200 MHz lower, all-around. But 400? :-O
What about the 14,16,18 versions then? This way 7980XE will have 1,5Ghz base clock
Unless OFC 7920x is meant to be somewhat similar to 7800x within the i9 line-up.
tamz_msc has the right idea, but didn't follow through. By comparing to existing Xeon chips, you can expect the base Skylake-X clocks to be very close to this:Have to say i am surprised, did not expect it to be under 3GHz. I anticipated very close clocks to 7900x, maybe 100-200 MHz lower, all-around. But 400? :-O
What about the 14,16,18 versions then? This way 7980XE will have 1,5Ghz base clock .
Why the 6150 with 165 W of power available to it and a much, much higher price? Why not compare it to the Xeon Gold 6140 (18 cores) with 140 W and a comparable price? That has a base of 2.3 GHz.http://ark.intel.com/products/120490
https://en.wikipedia.org/wiki/List_of_Intel_Xeon_microprocessors#Xeon_Gold_.28quad_processor.29
7980XE will have clocks on par or lower than Xeon God 6150 which has base 2.7 Ghz . All core turbo 3.4 Ghz . Max turbo 3.7 Ghz.
I think their 7800 and 7820x fit nicely with a aprox 330usd price. I think the best rationale must be to rake in the same revenue as prior gens. 14nm must be a good deal cheaper now so margins should not hurt much.Do you see any reason that Intel will vary much from the mid $300s price point that they have always targeted for this chip placement?
Top Intel non-HEDT desktop chips (x7xxK series):
7740X (June 2017): $339 - maybe could be lumped in with HEDT
7700K (Jan 2017): $350
6700K (Aug 2015): $339
5775C (June 2015): $366 - Slightly different chip than the rest (such as a much lower TDP), but figured I should put Broadwell in here
4790K (June 2014): $339
4770K (June 2013): $339
3770K (Apr 2012): $332
2700K (Oct 2011): $332
Why the 6150 with 165 W of power available to it and a much, much higher price? Why not compare it to the Xeon Gold 6140 (18 cores) with 140 W and a comparable price? That has a base of 2.3 GHz.
http://ark.intel.com/products/120485/
That make sense, I had missed it earlier and was going with 140 W as reported in most graphics and tables. If it is 165 W, then 2.6 GHz to 2.7 GHz base would be quite doable.http://www.anandtech.com/show/11464...ging-18core-hcc-silicon-to-consumers-for-1999
"Still covering the LCC core designs, the final processor in this stack is the Core i9-7920X. This processor will be coming out later in the year, likely during the summer, but it will be a 12-core processor on the same LGA2066 socket for $1199 (retail ~$1279), being part of the $100/core mantra. We are told that Intel is still validating the frequencies of this CPU to find a good balance of performance and power, although we understand that it might be 165W rather than 140W, as Intel’s pre-briefing explained that the whole X299 motherboard set should be ready to support 165W processors."
Coffee Lake tidbit of the day: It's a lot harder to pick a favorite from the new lineup than it was in the Kaby Lake generation, where Core i7-7700K is the clear choice for most people. Core i7-8700K is an achievement in its own right and a (very) worthy successor, but Core i5's performance is closer to the fastest SKU than it was before (based on Turbo).
Nah; I just need the top part (because I know what I need and want), so my question is, are they indeed going to be called 8700K etc.? And, is the 8700K the top SKU...?
And, because you know I have to ask (ha ha); how close is the turbo to the 4.2GHz leak we saw recently and in what respect (better or worse)?
Even more important than top turbo speed: what percent of the time can the chip stay at top turbo speed in intensive calculations? Intel could put a top turbo that is astronomically high but only allow it to be there for a short time. Or, Intel could be conservative with the top turbo and most people will assume desktop chips stay at top turbo indefinitely (not true with Kaby Lake, but not far off either).Better.
Even more important than top turbo speed: what percent of the time can the chip stay at top turbo speed in intensive calculations? Intel could put a top turbo that is astronomically high but only allow it to be there for a short time. Or, Intel could be conservative with the top turbo and most people will assume desktop chips stay at top turbo indefinitely (not true with Kaby Lake, but not far off either).
Correct.
Better.