Discussion in 'PC Industry' started by fellix, Oct 20, 2014.
Also note the lack of 2800X to better match Intel counterparts
Well Coffee Lake tops with the 8700K for now and the 2700X seems to be already pushing the envelope well enough through that 105W TDP.
I'm guessing the 1800X wasn't all that popular since most people could just buy the >30% cheaper Ryzen 1700 and get nearly the same results with some overclocking.
But the most popular was definitely the Ryzen 5 1600 and I'm guessing the 2600 will follow suit, so that's where AMD is really applying the pressure with the $200 price point.
I'm really hoping for Ryzen 2's CCX to come with 6 cores and 12MB cache now.
Still >5% single thread performance advantage for Intel is something
Hope to see some some reviews testing both Cofees and Ryzen 2 with memory kits to make them shine the most
So the clocks from above link
3.7Ghz base 4.35Ghz boost not that huge improvement vs 1800X 3.6/4.0 and with higher 105W TDP.
Looks like a much healthier bump vs 1700X 3.4/3.8 though.
I guess holding back on a 2800X assuming Intel brings out something quicker?
Higher TDP to allow it to hit/stay at the high boosts longer? (also with the newer turbo tech looking to keep higher turbo across more cores) Or because they're pushing hard against that performance cliff?
Clearly some improvement in that cliff though, gonna be interesting to see whether it will overclock much above 4.35Ghz. (what sort of OC is easy for Intel these days?)
Is better RAM speed -> uncore (or whatever AMD calls it) clock why L2/L3 seem quicker? (the CPU-Z screen says its at same CPU clock/RAM speed though)
Why do they have to have so many different turbo techs
Can't they just bundle it all together under one name?
I wonder sometimes if the different techs can wind up fighting against each other?
Nope, just cutting down the amount of models & matching Intel naming. They're also dropping Ryzen 3 x300X and Ryzen 5 x500X and moving to Intel-esque tick-tock-cycle where every other year comes new/improved architecture with new process and every other year process maturity + efficiency update to said architecture
Hmmm, the 2600X no longer matches clock (Base/Turbo) of the top CPU like the 1600X did. That makes it less attractive as an alternative. Where the 1600X you only sacrificed 2c/4t to the 1800x with the 2600X your sacrificing that plus base/turbo clocks compared to the 2700X.
Overclocking is still a thing.
Its not really intel-esque, its just a stepping revision for the enthusiast consumer market as 2 year cycle is a bit long for that market , notice no revisions for the APU or for EPYC. New steppings pushing performing forward bit by bit in the enthusiast market used to be a much bigger thing when intel/amd competition was very close P4, K7 days ( TB-A, TB-B, Barton).
I ended up having to get a 1700x and a b350 board but I am loving it and I am not to mad since I got the 1700x for $200 . I do hope they come out with a higher end APU perhaps a 6 core with a faster vega attached or even some hbm on die. I'm loving the 2400g system I bought my gf but would have liked something a little better. Hopefully next year they put something like that out and I can just upgrade hers
Yeah, I don't overclock the CPUs much anymore, the increased power consumption isn't worth it to me. For example, the 1600 OC'd to 1600x (stock) levels uses far more power than the 1600x at stock settings which made it a worse deal after a couple years (I usually use my CPUs for 2-4 years) depending on electricity costs in your area. The same was true for the 1700x versus the 1800x, except there the cost difference between the CPUs was large enough that you could still save a little bit of money over time with the 1700x.
Undervolting is where it's at
Preview stuff, not review till 19th.
Edit: hint at some special XFR sauce or a later 2800X?
There is no need for 2800X anymore with the new XFR. The chip ditches the old per core turbo range, it now adjusts clocks based on temps, vcore, current and other factors, if all parameters are good it will keep boosting even if all cores are fully loaded. So in effect, you could have the 2700X at 90% usage and still boosting to 4.1GHz (from 3.7GHz), as opposed to the 1800X which collapsed to the base clocks of 3.7GHz any time it's usage exceeded 2 cores.
So just hyping stuff we already know from the Vegan launch? https://videocardz.com/75194/amd-ryzen-2000-series-exposed-pricing-performance-leaked
Was hoping for something more than that.
Should be interesting to see performance with the new chipset since the top features need that & leaked/unsanctioned tests have all been on the older 370 chipset.
There is that remote possibility AMD will sanction few select top dies for halo overclockers product like 2800X Black Edition if it adds 100 or 200MHz headroom on top. Otherwise, as DavidGraham said, there is no reason for one to exist.
The launch slides are almost identical to the ones in the link above, 2700X is slower than a 8700K by 5% in 12 games @1080p with a GTX 1080. The link above has 50 games, I guess that's why the avg drops to -7.7%. In productivity apps, the 2700X is 20% faster than 8700K, while being cheaper by 40$.
I should have my 2700x on Thursday, will report back with results if I have some free time
Oh noes, i'll try to brace myself again for the pollution of this architecture thread with RAM settings and game run results ..
Are you getting a 470 chipset too?
There are SKU-specific threads for that purpose, and I for one welcome @Clukos's RAM benchmarks, as I'm eager to see whether (and to what degree) Ryzen remains sensitive to RAM settings now that its RAM and cache latencies have significantly improved, yet not to the point of matching Skylake's.