Discussion Ryzen 3000 series benchmark thread ** Open **

Page 10 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

zinfamous

No Lifer
Jul 12, 2006
110,514
29,100
146
These can't even beat an i5 9600K in gaming, and that's before overclocking. They OC like crap and need 1.4v to hit 4.3. Same old story as before. As soon as Intel gets their new node up and running, AMD will be right back in the moar cores bargain bin. I didn't want this to happen. I wanted AMD to run over Intel like a truck. No use in lying about the way I feel regarding these new CPUs. I wouldn't buy them for anything but maybe a dedicated video encoding rig or something and we already had threadripper for that, so...DOA IMO.

I predict video-documented consumption of a c.2021 Intel CPU box in your future.
 
  • Like
Reactions: john3850

exquisitechar

Senior member
Apr 18, 2017
655
862
136
Get ready for software fixes which will likely impact ST and light threaded performance.


View attachment 8143
This is like Zen all over again, glad I'm not an early adopter. It will take a while for the dust to settle so we can see the "real" performance of these chips, and which of the reviews are accurate. Different BIOSes, cooling, applied security mitigations, etc. all paint a very different picture.

Hopefully, AMD works on making their launches better in the future, since first impressions mean a lot. Thought that they would avoid stuff like this this time around.
 

zinfamous

No Lifer
Jul 12, 2006
110,514
29,100
146
This is like Zen all over again, glad I'm not an early adopter. It will take a while for the dust to settle so we can see the "real" performance of these chips, and which of the reviews are accurate. Different BIOSes, cooling, applied security mitigations, etc. all paint a very different picture.

Hopefully, AMD works on making their launches better in the future, since first impressions mean a lot. Thought that they would avoid stuff like this this time around.

Honestly, it's way better. How many months was it before a decent stock (not simply fully functional) MoBos were available for Zen 1? A lot more kinks out there for that release, and wildly inconsistent reviews from shop to shop, and just very real stock issues with inventory. The industry wasn't willing to bet on AMD at that time, and it showed.

It's kind of a big deal that this part has changed so quickly in 2 years, and it looks like these sort of BIOS releases and tweaks will be along much sooner. I mean, these results are already stunningly good compared to the competition.
 
  • Like
Reactions: lightmanek

Topweasel

Diamond Member
Oct 19, 2000
5,436
1,654
136
Honestly, it's way better. How many months was it before a decent stock (not simply fully functional) MoBos were available for Zen 1? A lot more kinks out there for that release, and wildly inconsistent reviews from shop to shop, and just very real stock issues with inventory. The industry wasn't willing to bet on AMD at that time, and it showed.

It's kind of a big deal that this part has changed so quickly in 2 years, and it looks like these sort of BIOS releases and tweaks will be along much sooner. I mean, these results are already stunningly good compared to the competition.
Yeah lets not confuse the usual early bios quirks with what happened with Ryzen 1k and the X58 (or was it the x79) platforms launches. As the margins lower in the difference the output is magnified making it the differences seem larger than they are. This is 2% here, -3% there type stuff, near margin of error. It does imply that there is a little more to be eeked out, but that doesn't mean there is a real problem with the platform at launch.
 

inf64

Diamond Member
Mar 11, 2011
3,686
3,958
136
It wouldn't be a real AMD launch if they would have no screw ups ;). Overall a superb uarchitecture, a true next gen Zen core. They beat Skylake IPC (if just by a few % on global average) and they only now lack the super high clocks of near 5Ghz that intel achieved on their multi year tweaked 14nm process. This is a game changer core and a balls of steel move by Lisa - going for ~15% IPC uplift with a chiplet design on a new node and without clock regression. Job well done!

I can't wait for Zen3, AMD MUST continue to aggressively pursue IPC as 7nm+ will offer close to zero clock advantages. I reckon they need around ~10% ST IPC to match Sunny Cove core on average. They will have decoupled core from the I/O, they can stack em up aggressively and beat intel in core/thread/power race. Let us hope TSMC does not screw up their 7nm/6nm nodes so AMD can execute like they did in the past 3 years - flawlessly.
 
  • Like
Reactions: lightmanek

exquisitechar

Senior member
Apr 18, 2017
655
862
136
Yeah lets not confuse the usual early bios quirks with what happened with Ryzen 1k and the X58 (or was it the x79) platforms launches. As the margins lower in the difference the output is magnified making it the differences seem larger than they are. This is 2% here, -3% there type stuff, near margin of error. It does imply that there is a little more to be eeked out, but that doesn't mean there is a real problem with the platform at launch.
I agree, it's nowhere near that (though I am seeing people with some issues), but I think this still matters for how the product is received. I'm more than satisfied with the performance, but just look at a few people's rather negative reactions here and elsewhere, mostly the gaming types - getting a few percent more performance across the board matters when the differences compared to the competition are small begin with. I'm seeing results that range from Ryzen trailing behind the competition by a solid bit to being within spitting distance in that particular area, *if* this update could have made more reviews show the latter, that would have been a good thing. Can't ever extract every drop of performance out of a chip at launch, I guess...
 

Mopetar

Diamond Member
Jan 31, 2011
7,797
5,899
136
Looks like this 7/7 release was rushed/too early for both Navi and Ryzen 3.

It doesn't matter if they pushed off the release until September, it would still be too rushed in one way or another. Having more time to get finished usually just amounts to more time to fart around on trivialities while the really tough problems remain unsolved.

I doubt we'd get 3% in games. Similar to MT results in rendering, we'd get nothing.

It might not do much for the overall framerates, but even 1% is big. As others pointed out when you're 3% behind, a 1% improvement shrinks the gap by 33%. I think it would probably do more to help the 1% or .1% frametimes which are just as important to some people as having a high FPS.
 

Asterox

Golden Member
May 15, 2012
1,026
1,775
136
Ubless you are deaf like me and can't get cationing to work. What does it say ? Please paraphrase

In short as conclusion or the very important detail he says, "who should still buy i7 8700K insted of Zen 2, and the answer is almoust none of you"!
 
Jul 24, 2017
93
25
61
  • Like
Reactions: turtile

Arkaign

Lifer
Oct 27, 2006
20,736
1,377
126
So I'm right in reading this that I'm not going to be meaningfully hampered on a 3700X if I keep my current DDR4-3000 CL15 Corsair Vengeance LPX sticks, right (for 1440p gaming with a card that's not a 2080 ti)?

Also, is the funny Zen quirk that runs 3000 MT/s memory at 2933 MT/s ironed out now?

Indeed. Hardware Unboxed also confirmed this. Between potato memory and 3600 low latency even, there was like 1% to see between it (in real world benches of productivity and gaming).
 
  • Like
Reactions: turtile
Feb 4, 2009
34,506
15,737
136
Indeed. Hardware Unboxed also confirmed this. Between potato memory and 3600 low latency even, there was like 1% to see between it (in real world benches of productivity and gaming).

Got a link for that? I’m particularly interested in potato memory vs high end memory.
 
  • Like
Reactions: turtile

CHADBOGA

Platinum Member
Mar 31, 2009
2,135
832
136
Indeed. Hardware Unboxed also confirmed this. Between potato memory and 3600 low latency even, there was like 1% to see between it (in real world benches of productivity and gaming).

TechPowerUp showed 8% in gaming, if you compared 2400Mhz memory to 3200Mhz memory.
 

lobz

Platinum Member
Feb 10, 2017
2,057
2,856
136
This is like Zen all over again, glad I'm not an early adopter. It will take a while for the dust to settle so we can see the "real" performance of these chips, and which of the reviews are accurate. Different BIOSes, cooling, applied security mitigations, etc. all paint a very different picture.

Hopefully, AMD works on making their launches better in the future, since first impressions mean a lot. Thought that they would avoid stuff like this this time around.
How is it like Zen all over again? Must some of you by all means be so hysterical?
 
  • Like
Reactions: kawi6rr

lobz

Platinum Member
Feb 10, 2017
2,057
2,856
136
This is what I was waiting for:

It's really funny, but honestly I don't understand how someone can say something like this.
the funniest part is, even random redditters know that improving the cache subsystem has one of the greatest effect on IPC - and this guy is supposed to be like the biggest professional of all times... honestly, he is basically juanrga with even worse English
 
  • Like
Reactions: CHADBOGA

Adul

Elite Member
Oct 9, 1999
32,999
44
91
danny.tangtam.com
no idea when the 12 core will be in stock so I said fuggit bought the 3600 for now. I plan to use that cpu for a plex encoder later.
 
  • Like
Reactions: IEC

Topweasel

Diamond Member
Oct 19, 2000
5,436
1,654
136
I agree, it's nowhere near that (though I am seeing people with some issues), but I think this still matters for how the product is received. I'm more than satisfied with the performance, but just look at a few people's rather negative reactions here and elsewhere, mostly the gaming types - getting a few percent more performance across the board matters when the differences compared to the competition are small begin with. I'm seeing results that range from Ryzen trailing behind the competition by a solid bit to being within spitting distance in that particular area, *if* this update could have made more reviews show the latter, that would have been a good thing. Can't ever extract every drop of performance out of a chip at launch, I guess...

The end is the key. I mean really think about it, outside Intel's refreshes that aren't even refreshes on the chipset side with later Z series releases. When is the usualy stopping point for bios updates for bug fixes. I think this is a little more transparent because we know when AMD and what version AMD is releasing in microcode on a very macro level. But the fact is most releases aren't perfect. Maybe nothing apparent, but never perfect and if the difference is at the high end PBO+XFR is missing out on an extra 100MHz or so, then it's been a pretty great launch.