AMD Bristol/Stoney Ridge Thread

Page 17 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Abwx

Lifer
Apr 2, 2011
10,847
3,297
136
Benchmarks!

seemingly DDR4 is @ 2133

http://www.bodnara.co.kr/bbs/article.html?num=134612&mn=4

PwkyYtI.jpg

That review is BS, how can a BR 4.2 score less than a FX 4.2 in CB ST..?..

cine-single.png


And how is it that the 3.9 7870K get about the same score than a 3.5 7800 :

cine-multi.png


Not even worth talking of the alleged 56W at idle of the 7870K in what looks more like a viral review than anything else, i mean whoever wants to do a serious review would check if his measurements and scores correlate with other reviews.

Result_IPC_CinebenchR15.jpg


So despite 4.2 they got lower score than the 3.8 X4 845 in ST and only 2.7% better in MT despite 8.5% higher frequency, if that s not a deliberately biaised review i dont know what it is...
 
Last edited:

Justinbaileyman

Golden Member
Aug 17, 2013
1,980
249
106
All this is, is a 845 with the IGPU activated on a AM4 Socketed PCB so it can use DDR4. Totally pointless in my eyes. I swear AMD better not me taping 4 of these cores together and calling it ZEN.
So far there is nothing Zen about this release.
 

NTMBK

Lifer
Nov 14, 2011
10,208
4,940
136
Core has had 256KB L2 since Nehalem. Just thought I'd point that out. Carrizo/BR's L2 cache performance is actually quite good. Its the lack of an L3 cache that's holding it back, not the size of the L2.

Bristol Ridge's cache setup is more like the old Core 2 Quad. The Q6600 had 2x4MB of L2 and no L3, for example.
 

Insert_Nickname

Diamond Member
May 6, 2012
4,971
1,691
136
Bristol Ridge's cache setup is more like the old Core 2 Quad. The Q6600 had 2x4MB of L2 and no L3, for example.

That was one of the nice things about the C2Q's, especially Yorkfield (2x6MB), a massive L2 cache (for the time). It also masked the memory latency very well.
 

Abwx

Lifer
Apr 2, 2011
10,847
3,297
136
One would have thought that Stoney Ridge was a 250-350$ laptops dedicated chip, but seems that HP has another opinion and they use it in their ProBook line, prices are quite high and most amazing is that for a hefty 50$ more you can have the same laptop but with a 2.4GHz Bristol Ridge...

http://www.notebookcheck.com/HP-Neue-ProBook-400-Serie-vorgestellt.174951.0.html

The ProBook 455 G4 is the only model to feature a processor made by AMD, being powered by a 7th generation APU with Radeon R4 or Radeon R6 graphics (three processor choices available). HP promises up to 16 hours of autonomy for these new machines

HP ProBook 455 G4 - 15.6-inch HD (full HD option available), 4.58 pounds, $499 USD
 

LTC8K6

Lifer
Mar 10, 2004
28,520
1,575
126
Well, the performance boost needs to be enough to justify replacing your motherboard.
 

Justinbaileyman

Golden Member
Aug 17, 2013
1,980
249
106
Ok well now you have my attention. I thought that there wasn't going to be any overclocking with these chips. Can we please see some benchies with this running at that 4.8Ghz?? Or is it not stable at those clock and was this just a Death Run??
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
Surprising to see that these appear to be unlocked, since the SKU doesn't indicate that.
Would love to know what the actual voltage was during the 4.8GHz validation, since CPU-Z is showing the default voltage for P0 PState (3.8GHz).
1.6V++ I would imagine.
 

Justinbaileyman

Golden Member
Aug 17, 2013
1,980
249
106
if its really 1.325v then these should easily hit 5.0Ghz unless CPUZ just isn't reading the Voltages correctly.
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
Obviously it is reading it wrong (new LPC/IO, on a new motherboard). If it was really able to do 4.8GHz @ 1.325V then why bother with a 4.8GHz validation in the first place, when it could certainly do much higher at higher voltage. When CPU-Z doesn't support the environment controller (LPC/IO), it reads the VID from CPU registers instead. When you overclock you are adjusting the multiplier of P0 PState and 1.325V is a normal VID for P0 on 9800.
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
If the default voltage for A12-9800 at 4.2GHz is 1.475V - 1.525V, I don't find the =< 1.6V to be too far fetched.
 

Abwx

Lifer
Apr 2, 2011
10,847
3,297
136
1.6v on air ??

If the corean site has at least his CPUZ shot right then it should be anything in the 1.5-1.55V range.

Since this APU is clocked up to 4.2 at stock it s no wonder that it should have at least 10% frequency margin left, here that would make 14% and from AMD published curves we know that BR use a process that has at least 10% frequency uplift above Carrizo s.

dvix_9800.jpg
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
It has a "D" letter above of the "made in / diffused" in text, so it is EVT silicon shipped to the partners / press.
Marked as a retail part, but still evaluation silicon.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
2 x 1MB L2?? this is getting worse and worse :mad:

It's about what we expected though. BR is just Carrizo with some minor tweaks.

The only reason I can see them doing that is so as not to cannibalize Zen sales.

The reason to do it is to save money. They didn't want to do any kind of redesign on an XV-based core when all their eggs are in the Zen basket. AMD still has limited resources. It was easiest for them to take the work they had done on Carrizo and drag-drop it into Bristol Ridge to serve as a stopgap until they can launch Raven Ridge. There's no way a 2x2Mb l2 cache would cause Bristol Ridge to cannibalize Summit Ridge sales.

Wait to see more about different DRAM config comparison in game, since DDR4 could vary from 2133 to 3600Mhz......

Don't expect DDR4 speeds above 2400 MHz out of BR unless bclk OC is easy and convenient on AM4.


I would like to know more about that setup. Like his memory speed/timings and some other things. Looks pretty good, though. For anything that doesn't need the extra l2, that 4.8 GHz BR ought to smash Kaveri.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Exactly, that's what Carrizo has. Kaveri/Godavari is 2x2MB, Carrizo/Bristol Ridge is 2x1MB.

Carrizo/Bristol Ridge have significantly faster L2, and in a majority of non-gaming benchmarks, the XV-based CPUs whip the SR-based CPUs soundly. In a fair number of games, the opposite is true.
 

Shehriazad

Senior member
Nov 3, 2014
555
2
46
Carrizo/Bristol Ridge have significantly faster L2, and in a majority of non-gaming benchmarks, the XV-based CPUs whip the SR-based CPUs soundly. In a fair number of games, the opposite is true.

Well most of the gaming comparisons weren't "normalized" to the same mhz.

3.5 Ghz XV base clock vs 3.7/3.9/4.0 Ghz....hardly seemed professional.

I do remember that the total average mhz for mhz difference was like 2-3% in gaming for Kaveri/Godavari if I recall correctly.

Though I'm not sure this will still hold up...aside from that XV on AM4 will sit at least on 3.8 Ghz base for their "top dog" for now...might even be higher, not like we already know the final specs of the DIY Chips. (960(K),970(K) and so on)

In addition to that we get different ram and updated Chipset...these should all play in Bristol Ridges' favor and at least remove those pesky 3% average hz for hz gaming "loss".



Not sure how much This wccftech article is to be trusted...but at least there it looks like that even the shitty OEM version seems to outperform Kaveri on all ends. But then again...the test there is also hella awkward. 1600mhz DDR3 vs 2133mhz SINGLE CHANNEL DDR4...just a big plain "wat?" xD