Read more.Quote:
Coffee Lake launch is oddly timed, to coincide with "the Great American Solar Eclipse".
Printable View
Read more.Quote:
Coffee Lake launch is oddly timed, to coincide with "the Great American Solar Eclipse".
Intel's management is woefully inadequate to the new task in hand of actually having a competitor who can match or beat there multi-threaded performance at a massive cost saving. Not to mention that AMD have announced that AM4 will support Ryzen processors for at least three refreshes. Coffee Lake is coming out on another chipset so isn't compatible with 100/200 boards. If you're a monopoly for so long you end up a dinosaur and can't react quick enough and you end up dying a long and painful death. Intel rules no more thankfully there is a new King on the horizon.
Dunno Intel will come back in the game again eventually.. whether if it is Intel or AMD who has the crown doesn't matter too much to me, but I whichever at current point is good and seem solid enough at the given time it is available... then I buy when I see a need for an upgrade.
Hopefully the people that have laptops with a 7700hq won't be that upset as this is a bit fast! Intel suck and only care about the $$$$.
Isn't it about time that people finally understood the difference of there, their and they're?!
What this will potentially cause is a massive shake up in management as investors look at what is happening with the brand loyalty being shaken because of what AMD has done. If they realise what is happening then they will pressure Intel to change their ways because if they don't, they will lose investment.
I think we are likely to see a Lisa Su-esque shame up of Intels CPU management and r&d staff so that Intel is no longer dragging their feet.
I know we all think that Intel probably has a few brand new architectures sitting in the closet but I actually think they don't. This is because Intel are an investor based company so if Intel is found to be wasting large sums of money for no feasible near future reward then it will be shut down. So their may be plans but no further. In which case Intel could be years off of a new uArch that moves them away from the dusty burlap sack that is the Core series. Don't get me wrong, the core series is fantastic but for 8 years it has just been stretched further and further which now makes an "enough is enough" statement.
Intel is at the stage AMD was a few years ago. Outdated architecture in some ways but great in others. The pressure has all been on AMD so let's see what the heat on Intel can do
"I think we are likely to see a Lisa Su-esque shame up of Intels CPU management and r&d staff"
Nah. Intel knows EXACTLY what it's doing. "Let the peasants scrabble in the dust feeding on this AMD release... ". Ideally they should have, at least, the next three major releases planned and ready to go. NO company is ever going to release their fastest, bestest stuff, jumping forward too far as that's not good for $$$$. Intel has been drip feeding consumers for years.
It's time to buy AMD just to teach Intel a lesson but, as soon as Intel releases their new stuff we'll all just but Intel again.
There has been a leak of the Core i7 8700 performance in the CPU-Z benchmark.
The 1.78.1 version can be downloaded here:
http://www.filepuma.com/download/cpu...3419/download/
My Xeon E3 1230 V2/Core i7 3770 at 3.7GHZ gives a score of 1538 and the Core i7 7700K score is from the reference list in CPU-Z.
http://i.imgur.com/4ThCpYD.png
The leak:
http://i.imgur.com/857LuQp.jpg
If you assume the MT score is at 3.7GHZ for the Core i7 8700K,if you were to scale it upto 4.2GHZ but 4 cores instead it will give you roughly 10600 points,or around a 6% IPC increase.
Now assume the ST score is at 4.3GHZ and the Core i7 7700K is at 4.5GHZ,if you scale the Core i7 8700K score upto 4.5GHZ,it actually also gives you roughly a 6% IPC increase,which matches the MT score.
interesting bit from the Doc from Anandtech was the fact that both of the i3's support ECC
It's the Pentium III vs Athlon .... Athlon to Core cycle all over again
So far (over say the last decade), Intel's management have gotten away with wasting money on the following:
- Contra revenue (aka dumping) to support Atom after neglecting it for years to the tune of $4 billion
- Larrabee to the tune of around $1billion (although some if this lives on in Xeon Phi)
- McAfee for $7.6 billion.
- Infineon Wireless for $1.4 billion
- Altera for $16.7 billion
- MobilEye for $15 billion
Now, some of these things have made some profit but a lot (most?) have not. So spending/wasting money is something Intel is good at. Guess mainly because they are so profitable that they can afford to do these things even when they don't work out.
EDIT: what I meant to say is that the feasible near future reward for all of the above was at the very least very risky.
Since it's the same architecture as Skylake and Kabylake, surely any IPC increase can only come from more cache or similar?
Hm, obviously wait for the release where usually hardware.fr can usually relied on to compare it at the same clocks, but I was under the impression that Coffee Lake is another optimization of Kaby Lake (which in turn was n optimization of Skylake).
The wikipedia calls it a "2nd Optimization" in the forward link from Kaby Lake.
So while tick-tock seems to be officially dead and has been replaced by process-architecture-optimization, it seems that still doesn't tell the full story as for 14nm it looks more like this:
process: Broadwell (but aside from new graphics, didn't that also add some new instructions?)
architecture: Skylake
optimization: Kaby Lake
optimization 2: Coffee Lake