AMD Carrizo Pre-release thread

Page 117 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

monstercameron

Diamond Member
Feb 12, 2013
3,818
1
0
The designs are way too different for any kind of proper comparison. I haven't worked too much on recent Intel designs, but I would say AMD has better power management than the most recent Intel designs do. That's the result of being so far behind in both the process technology and the design itself, however that's definitely not a bad thing.

Once AMD has access to smaller manufacturing process (such as the 14nm LPP) their power efficiency will be pretty hard to compete with. Let's just hope the new architechture lives up to the expectations.

While the power management in Carrizo for example is extremely advanced and effective, there are still some things which need to be improved heavily. Basically none of the features of the power management are documented, let alone properly. Because the lack of documentation, the system manufacturers are unable to tune the power management properly to provide the best possible performance on their system.

I have personally spent maybe < 800 hours on the SMU, PMU and PMC on Steamroller and Excavator. Without the breakthroughs I made on Steamroller (i.e learnt how to program and configure the SMU) I couldn't change a single setting either on Steamroller or Excavator, despite having access to the same documents as the manufacturers do. Most of the time the stuff just isn't documented at all and when it is, it is done either poorly or just wrong.

Are these documents usually under NDA or are they open like the gcn isa docs?
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
Are these documents usually under NDA or are they open like the gcn isa docs?

The public documents are somewhat stripped, however what comes in configuring the power management or programming the system management unit (SMU) the NDA and the public contain the same information (none) D:
 

FlanK3r

Senior member
Sep 15, 2009
312
37
91
No rush, but its very great CPU for SUperpi. Mr Chew* or mr Stilt must to find 70 dollars and buy this CPU for late evenings :up:



I had IDE mode and the issue of this piece was IDE BCLK black screen. I tried another ways as lower RAM, lower APUNB or higher voltages at SB, VDDNB, CPU and doesnt change the issue. So this was my highest stable BCLK. I believe piece of piece of 845 Athlon will be little different. Maybe PCIe card slot with M2 could be solution (into the last PCIe x4 slot). Friend told me, old HDD raptor is more SATA stable also. I used SSD.
Seems CPU si scaling very well. Next time I must tweak the OS or to use XP for SUperpi. Clear 10 minutes is with these CPU/CPUNB clocks doable. Interesting, for this CPUNB was OK only around 1.17V! I dont tested, if CPU has coldbug or not. With few liters of LN2 I left around -50 C temperatures.

 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
So that was an LN2 shot, makes sense now.

btw you don't need to go as expensive as NVMe harddrives or whatever, you can just grab a cheap PCIe SATA controller and try that. Though some of those get cranky at higher PCIe speeds. The cheap-arse one I bought (that has since burned out) was good for bclk up to 119 MHz.
 

el etro

Golden Member
Jul 21, 2013
1,581
14
81
It looks like it can't sustain 4Ghz on all cores being into 65W TDP.

Anyway will be a blast if efficiency goes up at this performance levels.
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
It looks like it can't sustain 4Ghz on all cores being into 65W TDP.

On Carrizo holding 3.4GHz on both compute units requires around 50W TDP limit, in workloads such as Cinebench R15.

I would expect Bristol Ridge to hold around 3.7GHz on both units with it's 65W TDP limit. In combined workloads (CPU+GPU) the CPU frequency should drop to 3.1GHz region.
 

Flash831

Member
Aug 10, 2015
60
3
71
If I have understood correctly, Kaveri is being produced on some custom SHP 28nm silicon, whereas Carrizo/Bristol Ridge is being produced on 28nm bulk silicon,.

As AMD will soon introduce the AM4 platform, Bristol Ridge will become the mainstream CPU for the platform and Kaveri should therefore be End-of-life.

Asuming the transition goes as planned, does anyone have a clue how much savings AMD makes in their Bill-of-materials switching their production from SHP to bulk silicon? :sneaky:
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
Steamroller (Kaveri / Godavari) is / was made on GF28A, which is a tuned (custom) version of the standard GF28HPP. They are both bulk nodes, since GlobalFoundries cancelled their plans for SOI 28nm process (GF28SHP).

Carrizo / Bristol Ridge & Stoney Ridge are manufactured on the standard GF28HPP process.
 

Flash831

Member
Aug 10, 2015
60
3
71
Steamroller (Kaveri / Godavari) is / was made on GF28A, which is a tuned (custom) version of the standard GF28HPP. They are both bulk nodes, since GlobalFoundries cancelled their plans for SOI 28nm process (GF28SHP).

Carrizo / Bristol Ridge & Stoney Ridge are manufactured on the standard GF28HPP process.

I understand if you can't answer this, but wouldn't it be cheaper to produce chips using the standard GF28HPP process than on the custom version?
 

The Stilt

Golden Member
Dec 5, 2015
1,709
3,057
106
I have no idea what the price difference was, to be honest. I would imagine that GlobalFoundries felt pressure about increasing the prices for GF28A parts due the already low and still declining demand. Afterall only a single design was ever manufactured using the process so having a separate production line for a low volume product is not necessarily too cost effective.
 

NostaSeronx

Diamond Member
Sep 18, 2011
3,683
1,218
136
I have no idea what the price difference was, to be honest. I would imagine that GlobalFoundries felt pressure about increasing the prices for GF28A parts due the already low and still declining demand. Afterall only a single design was ever manufactured using the process so having a separate production line for a low volume product is not necessarily too cost effective.
GF28A is not the complete node designation.

Qualcomm; GF28LPQ / GF28Q
AMD; GF28LPA / GF28A

LPQ = Qualcomm optimized node
LPA = AMD optimized node

The LP nodes at GlobalFoundries are copy-smart versions of SEC28LPH. SLP/HPP are pure GlobalFoundries nodes.
 

monstercameron

Diamond Member
Feb 12, 2013
3,818
1
0
WE'RE BACK BOISSSS!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
http://www.bestbuy.com/site/hp-envy...k-silver/5295100.p?id=bb5295100&skuId=5295100
5295100_rd.jpg;canvasHeight=550;canvasWidth=642

5.6" Full HD multitouch screen for hands-on control
The 1920 x 1080 resolution boasts impressive color and clarity. Touch, tap, glide and make the most of Windows 10. IPS technology. WLED backlight.
8GB system memory for advanced multitasking
Substantial high-bandwidth RAM to smoothly run your games and photo- and video-editing applications, as well as multiple programs and browser tabs all at once.
1TB hard drive for ample file storage space
Holds a sizeable collection of digital photos, music, HD videos and DVD-quality movies. 7200 rpm spindle speed enables rapid read/write times for handling large files and complex applications.
360° flip-and-fold design
Offers versatile functionality with laptop, audience, tabletop, presentation and tablet modes.
FreeSync technology for fluid gameplay
AMD FreeSync synchronizes the refresh rate between the GPU and display to eliminate image tears and choppiness.
Cloud support lets you access your files anywhere
Store your photos, videos, documents and other files on Dropbox for secure access across multiple devices. Fees may apply.
AMD Radeon R7
Integrated graphics
chipset with shared video memory provides solid image quality for Internet use, movies, basic photo editing and casual gaming.

Time for a new laptop! stand aside skull canyon!