AtenRa
Lifer
- Feb 2, 2009
- 14,001
- 3,357
- 136
Cheapest EU prices
http://geizhals.eu/?fs=core+i7+7700k&in=
Core i7 7700K = 361 Euro
Core i7 6800K = 435 Euro
Core i7 6900K = 1100 Euro
8370E shipped 2 years later than 8350 and was indeed a new stepping (Vishera-K), which itself shipped 1 year and something later than 8150, which was one of their first CPUs on their at-the-time new node, 32nm SOI.
Your comparison doesn't stand really.
32nm had a better clock/vcore/pwconsumption curve at the clocks we are talking about than 28nm, so that leaves 1 node less. The other one I can think is 20nm, and then 14nm LPP that mixes parameters from their 20nm and a true 14nm node.
We have seen a 3.4ghz 8c/16t consuming just a bit higher than 95w on blender/handbrake, with a little overvolt to assure completely stability on a live event. Suddenly the process improved so much or the binning is that aggresive (we are talking about AMD here) that now 3.7ghz all-core turbo or base at 65w is realistic at launch? Come on, blue tinted glasses are as bad as red tinted ones in my book.
Pro tip: The article implies they are talking about BASE clocks because they mention that turbo clocks aren't mentioned (so we can only think of an mild all-core turbo, or just base clocks).
If true, that regular 1700 is my next CPU purchase. But its a bit less aggressive then I expected to be honest, but it seems fair enough. ~€400 for a 65W 3.4(?)/3.7GHz 8C/16T CPU is quite impressive on multiple levels, and is a worthy upgrade from my 3770non-K...
SAAA
Senior member
- May 14, 2014
- 541
- 126
- 116
You keep making comparisons to FX line in Zen threads but I don't think 8c FX are any similar to 8c Zen, short explanation:
FX quad module transistor count = 1.2 billions (for the whole 315mm2 cpu)
Zen quad cluster transistor count = 1.4 billions (ISSCC paper, 44mm2 CCX only)
You can easily see that Zen eight cores have more than twice the transistors, if that's true I doubt they could keep less than 65 watt under load even at iso frequency: 14nm and finfets help but 2x resources at 0.7x the power is pushing it way too much.
Burpo
Diamond Member
- Sep 10, 2013
- 4,223
- 473
- 126
@Insert_Nickname
Depends on what you're doing with it. For games I doubt you'll see much improvement over your 3.4 Ivy..
Rendering or other heavily threaded tasks will of course be a different story..
Depends on what you're doing with it. For games I doubt you'll see much improvement over your 3.4 Ivy..
Rendering or other heavily threaded tasks will of course be a different story..
Is it a coincidence that wccftech posts about linley group from last august now?
http://wccftech.com/ryzen-smaller-die-intel-zen-architecture-not-good-hpc/
http://wccftech.com/ryzen-smaller-die-intel-zen-architecture-not-good-hpc/
You hit the nail on the head there. Combine with video encoding and occasional dabbling in multiple VMs, and you can see why such a CPU seems a good fit. Gaming is a secondary concern, since I'm not CPU limited currently at 1440p anyway.
The low 65W TDP is just icing on the cake. For comparison my current 3770 uses almost exactly 77W overclocked to 4.3GHz, stock is around 60W.
coercitiv
Diamond Member
- Jan 24, 2014
- 6,185
- 11,845
- 136
Go with the current parts and let Zen settle in, both from a process maturity and pricing. And lose all the fun of testing Zen
David Kanter confirmed my theory that Zen has same FO4 than BD, so the matter is only the number of transistors. 8 zen cores so have double the transistors than BD, but the 14nm LPP has up to 60% less power consumption. Consider all the energy saving techniques that are in BR and will be in Zen and there are not in the FX8350 and you will see that we are not so far from the truth.
Moreover a 3.6GHz 95W sample exist. With cubic scaling 3.3GHz are 72W... But the 3.6GHz is an ES... With a little binning a 3.3GHz 65W can be feasible... Finally I once calculated 3GHz 45W...
Glo.
Diamond Member
- Apr 25, 2015
- 5,704
- 4,548
- 136
And you know what was most shocking? Keller was talking about a lot of hardware that ultimately has appeared in Zen and to some degree with Vega. Its from 2014.
I am wondering what will happen with K12, because I am 100% sure, that closing whole project would be complete and utter waste of work they put in it. And if you think about that Keller, was one of the people that helped design Apple MacroScalar architecture(A6), and compare what Apple was able to build upon it today...
The imagination starts to wonder. But that is a topic for another post and another thread.
CatMerc
Golden Member
- Jul 16, 2016
- 1,114
- 1,149
- 136
uop cache let decoders to turn off, for instance... Even in prime 95... Then there is the stack/memfile that let avoid to use power hungry ALUs for stack management, then there is move elimination, that let not use power hungry ALUs for simple moves, moreover there are more dense uops, that require less energy to be managed, finally there is more aggressive clock gating...
Even in prime 95 there is margin for improvements...
One thing that people tend to forget is that ryzen was demoed against i7 6900K in a game. It was Battlefield 1 and it really held its own. That makes me hope that gaming benchmarks wont be too bad for Ryzen.
You once 'calculated' 4GHz base 95W LAUNCH too... With 4.5GHz Turbo. Based on your FO4+Neon claims.
Say so a billion wrongs things and you're bound to get something right.
Sent from HTC 10
(Opinions are own)
TRENDING THREADS
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 9K
-
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 7K
-
Facebook
Twitter