ASRock FM2A88X Extreme6+ / A10-7700K observations

[justapost]

Having an A10-770K running in an ASRock FM2A88X Extreme6+ for a few days now, so it's time to share banchmarks and observations. Did alot of linux testings and benchmarks. Over at phoronix there was an Kaveri review so i ran my tests against these.


All results beside 7700K are from the Phoronix review. I grouped em in single and multi core benchmarks. The native ones where build with special optimisations on the phoronix review, i did not yet figure out how this was done, so mine where build with default compiler settings and are not directly comparable.

My Specs:
CPU: A10-7700K
BOARD: ASRock FM2A88X Extreme6+
RAM: 2x4GB Crucial Ballistix Tactical DDR3-1866, CL9-9-9-27 (XMP Profile).
SSD. Samsung 840 120GB

Air Cooling: AMD PII Black Edition 125W Heatpipe Cooler
Water Cooling: D-Tec Fusion v2 / Swiftech MCR320 (3x120mm) / Laing D5-Vario 1/2 AG / DangerDen 5,25" Dual Bay Reservoir.

There are two usefull benchmarks for stability testing under linux. The "Open Porous Media" benchmark turned out to be very usefull for quick stability testings and required the highes voltage to pass. Second, the "Primesieve" is the prime95 linux pondon, generates alot of heat is the most energy consuming.

Here is an summary of the lowest voltage required for "Open Porous Media" on various voltages along with the benchmark results and the power consumption at the wall. The turbo mode does not work on that board as soon as mutipliers and voltages are adjusted manually even with C6 and APM enabled. CNQ was enabled.

Freq	VID	Idle	Load	Result	Power
3400	1,1625	58	117	96,58	11300
3500	1,1625	58	117	95,36	11157
3600	1,1625	58	117	94,38	11042
3700	1,1625	58	117	93,09	10891
3800	1,1625	58	118	92,48	10912
3900	1,18125	58	120	92,28	11074
4000	1,2	59	126	91,13	11482
4100	1,2375	59	135	91,27	12321
4200	1,28125	59	144	90,33	13008
4300	1,33125	59	158	89,35	14117
4400	1,425	59	185	89,37	16533
4500	1,49375	59	202	90,47	18275

At around 4400MHz the cpu started to throttle, this happens if the onchip radeon temps exceed 65C. The throttle effect is also visible at the multi core results above at 4.4GHz comparing air and water results. Switching to water cooling reduced the radeon temps from 65C down to 40C at 4.4GHz. At 4.6GHz however they reached 65C again running the primesieve benchmark and the system usually froze.
Will receive an A8-6500 later today so i can run a few benchmarks to get clock vs. clock and turbo efficiency results.

[justapost]

Received the A8-6500 and ran the linux benchmarks at stock (3.5/4.1GHz) and at 3.4GHz for comparison. Heres an link to the results http://openbenchmarking.org/result/1...SO-1401211SO29. Running a few windows benchmarks at the moment for an clock vs clock comparison. Feel free to request benchmarks you are interested in.

[Lightman]

Excellent voltage/stability table! Thanks!

I received my new A10 7850k today but came back too late from work to pick it up from my office. Will have to wait till tomorrow for a chance to play with it

Regarding testing, can you please test Frybench as it is quite sensitive not only to CPU speed but also cache and memory. Few of the older benchmarks like real time ray-trace engine AraunaBench and maybe Cinebench 2003 x64 as I have quite a few results to compare to. Other than that whatever games you prefer, ideally @720p and 1080p medium to high details

Thanks!

[justapost]

Excellent voltage/stability table! Thanks!

I received my new A10 7850k today but came back too late from work to pick it up from my office. Will have to wait till tomorrow for a chance to play with it

Regarding testing, can you please test Frybench as it is quite sensitive not only to CPU speed but also cache and memory. Few of the older benchmarks like real time ray-trace engine AraunaBench and maybe Cinebench 2003 x64 as I have quite a few results to compare to. Other than that whatever games you prefer, ideally @720p and 1080p medium to high details

Thanks!

Hmm somehow my earlier reply got lost. Thank you for the feedback and benchmark recommendations. It's an interesting new tool hope you enjoy it as well. Looking forward to your results on the mATX version, because i need an solid one for usual office builds. Having an GBT F2A88XM-HD3 here which struggles running prime95 with the A8-6500 at stock.
So far i grabbed frybench and cb 8.5. AraunaBench is only available as v2 which requires cuda. Can you upload the older version somewhere online temporary, seems to be an interesting benchmark.
Gaming benchmarks will be difficult because i do not own a single one.

[justapost]

Richland - Kaveri comparison clock vs. clock

Richland:
A8-6500 CPU 3.4GHz NB 1.6GHz MEM 1866MHz CL9 GPU 4 Units at 800MHz


Kaveri:
A10-7700K CPU 3.4GHz NB 1.6GHz MEM 1866MHz CL9 GPU 6 Units at 800MHz


Benchmarks:

Benchmark	Improvements	Post with Details
3DMark 2013:	30%	#14
7Zip:	5%	#15
As SSD		#16
Aida64	9%	#17
Arauna	3%	#18
Cinebench	2% (single) 8% (multi) 7% (OpenGL)	#19
CompuBench	12% (cpu) 216% (OpenCL)	#20
CPUMark	5%	#21
Crysis Timedemo	12%	#22
FritzChess	7%	#23
Frybench	0%	#24
Geekbench	9% (single) 13% (multi)	#25
Luxmark	-3% (CPU) 165% (GPU) 75% (CPU+GPU)	#26
Nuclearus MC	8% (single) 27% (multi) 5% (memory)	#27
PCMark 8	1%	#28
ratCPU	286% (CPU) 213% (GPU)	#29
ScienceMark 2	5%	#30
Solidworks 2013	5% (cpu) 7% (gfx)	#31
Specvieperf	-7% (v10,single,1280x1024) -16%(v10,multi,1280x1024) 22% (v12,single,1900x1060)	#32
Sunspider Javascript	4%	#33
Truecrypt	25%	#34
Unigine Heaven	69%	#35
Wprime	40%	#36

[justapost]

While i wait for the richland benchmarks to finish i tried an adata 2600 kit.
2500MHz CL13-13-13 stable on an quick test, above i guess the BCLK is causing trouble. Kit is rated for 1.65V had to use 1.675V for those clocks.

[Lightman]

Hmm somehow my earlier reply got lost. Thank you for the feedback and benchmark recommendations. It's an interesting new tool hope you enjoy it as well. Looking forward to your results on the mATX version, because i need an solid one for usual office builds. Having an GBT F2A88XM-HD3 here which struggles running prime95 with the A8-6500 at stock.
So far i grabbed frybench and cb 8.5. AraunaBench is only available as v2 which requires cuda. Can you upload the older version somewhere online temporary, seems to be an interesting benchmark.
Gaming benchmarks will be difficult because i do not own a single one.

My Cinema4D with database of results:
https://www.dropbox.com/s/c70dd02dyt...nch2003x64.zip

AraunaBench:
https://www.dropbox.com/s/bow88qmugo...raunaBench.zip


Great work so far!

[Daveburt714]

Having an A10-770K running in an ASRock FM2A88X Extreme6+ for a few days now, so it's time to share banchmarks and observations. Did alot of linux testings and benchmarks. Over at phoronix there was an Kaveri review so i ran my tests against these..

Hi Just,
Thanks for all your tests!

I noticed you changed the mobo to a Gigabyte... Were there problems with the ASRock EX6+?
I've grown fonder of ASR boards, but they do seem to be a little slow with bios updates compared to Asus...
It's been ridiculously cold here (-7f right now ) and it looks like the cold temps are going to stick around for a while.

I've been jonesing for some benching fun, and I had plans to get a 7700k for the HTPC anyway. I've already bought/upgraded the HTPC to an FM2+ ASRock Pro 4+ in preparation, but the cold weather is making me think I might just get a new ATX board with it for some cold air/water fun!

If the EX6+ was a little finicky, I'll just go with the Asus A88X Pro. MicroCenter has some good deals on Kaveri bundles...
http://www.microcenter.com/site/prod...d_bundles.aspx

For $20 more, I may just go ahead and get a 7850k... I can always just underclock/volt it for the HTPC when I'm done playing...

Thanks,
Dave

[justapost]

My Cinema4D with database of results:
https://www.dropbox.com/s/c70dd02dyt...nch2003x64.zip

AraunaBench:
https://www.dropbox.com/s/bow88qmugo...raunaBench.zip


Great work so far!

Great, thank you Lightman! That list is extensive. Benchmark came just in time, runninge Spec10 atm afterwards wprime, now also araunabench and i'm finished with richland.

[justapost]

Hi Just,
Thanks for all your tests!

I noticed you changed the mobo to a Gigabyte... Were there problems with the ASRock EX6+?
I've grown fonder of ASR boards, but they do seem to be a little slow with bios updates compared to Asus...
It's been ridiculously cold here (-7f right now ) and it looks like the cold temps are going to stick around for a while.

Hi Dave,

There where no problems with the EX6+ beside it was bussy with richland benchmarks. Had an gbt matx fm2+ also lying around here, also an new memory kit so i assembled an quick setup for memeory testings. Had a few troubles with that gbt board earlier but it turned out the just replaced psu was buggy.

I've been jonesing for some benching fun, and I had plans to get a 7700k for the HTPC anyway. I've already bought/upgraded the HTPC to an FM2+ ASRock Pro 4+ in preparation, but the cold weather is making me think I might just get a new ATX board with it for some cold air/water fun!

If the EX6+ was a little finicky, I'll just go with the Asus A88X Pro. MicroCenter has some good deals on Kaveri bundles...
http://www.microcenter.com/site/prod...d_bundles.aspx

For $20 more, I may just go ahead and get a 7850k... I can always just underclock/volt it for the HTPC when I'm done playing...

Thanks,
Dave

Bought the 7700K because i do not need much igfx power. According to the AMD slides all kaveri chips support cTDP, so i thought i can run mine at 65W or 45W after i'm done with benching and overclocking. Expected the 7700k would require abit less in idle because of the lower count of gpu units. Thing is i can select 45/65W as cTDP in the bios but it has no effect.
The EX6+ allows voltage offsets for cpu/NB, eighter chip internal or on the pwm, so it's prepared for underclocking/volting. The GBt board has no such features. Grabbed the EX6+ because it has an hdmi input to be used to pass other devices thru. Also it has an additional USB3 chip and an better sound chip. Power consumption is 5-10W higher in idle compared to the gbt, guess due to all the additional features and extra slots.

[Lightman]

Yes, my A10-7850K supports cTDP and now I'm running 45W CPU in my AsRock board. BTW AsRock just released new BIOS for it, ver. 2.00 and it feels solid! So far no problems, but I'm mainly underclocking and fooling around today. One feature I really like is build in BIOS Flash utility which works even without an USB drive or HDD. I've simply installed Richland CPU, added some RAM and booted system to BIOS, plugged network cable and 5 minutes later I had newest firmware installed ready for Kaveri!

@justapost - one of my favourite benches and all database entries were done on my hardware at different points in time Most of them properly described as well!

[justapost]

Yes, my A10-7850K supports cTDP and now I'm running 45W CPU in my AsRock board. BTW AsRock just released new BIOS for it, ver. 2.00 and it feels solid! So far no problems, but I'm mainly underclocking and fooling around today. One feature I really like is build in BIOS Flash utility which works even without an USB drive or HDD. I've simply installed Richland CPU, added some RAM and booted system to BIOS, plugged network cable and 5 minutes later I had newest firmware installed ready for Kaveri!

Just checked theres a new bios for the EX6+ available.

Changelog:
1. Modify share memory rule.
2. Update EZOC table for 7700K and 7850K.
3. Add "Target TDP" item.
cTDP had been there with 2.4 must try if it works. Now if they'd add downcore options for cpu and gpu i could run an better comparison against the A8-6500 with it's four gpu units. Good catch with bios update utility, had missed that and flashed via USB earlier.

@justapost - one of my favourite benches and all database entries were done on my hardware at different points in time Most of them properly described as well!

Consolidated all my local storage disks last year, so atleast i have collected all the spread around benchmark results in one folder now, but it's still an muddle.
Started benching the kaveri, will add results to my earlier post later but meanwhile here's cb2003 and arauna.

A8-6500 3400/1600/1866/800


A10-7700K 3400/1600/1866/800

[justapost]

Added results to #5

[justapost]

3DMark 2013:

Benchmark	A8-6500	A10-7700K
3D Mark Firestrike	843	1325	57%
3D Mark Cloud Gate	5172	6335	22%
3D Mark Ice Storm	55617	61638	11%
3D Mark 2013			30%

A8-6500


A10-7700K

[justapost]

7Zip:

Benchmark	A8-6500 Score/Usage	A10-7700K Score/Usage
7Zip Compression Score	2787	2912	4%
7Zip Decompression	2624	2790	6%
7Zip			5%

A8-6500


A10-7700K

[justapost]

As SSD
Kaverie scores are lower, maybe because of the partition beeing nearly full. Only ran it as an test for disk degradations anyway.

A8-6500


A10-7700K

[justapost]

Aida64

Benchmark	A8-6500	A10-7700K
CPU AES (MB/s)	7256	7531	4%
CPU Hash (MB/s)	1720	2387	39%
CPU Photo (MPixel/s)	8082	8591	6%
CPU Queen (Score)	17020	16775	-1%
CPU ZLib (MB/s)	145.1	153.1	6%
CPU			11%
FPU Julia (Score)	5673	5769	2%
FPU Mandel (Score)	2928	2960	1%
FPU Sin (Score)	1192	1328	11%
FPU VP8 (Score)	2833	3190	13%
FPU			7%
MEM Read (MB/s)	17751	18791	6%
MEM Write (MB/s)	8364	10240	22%
MEM Copy (MB/s)	14737	19043	29%
MEM Latency (ns)	77.2	98.6	-21%
MEM			9%
AIDA			9%

A8-6500
CPU

FPU

MEM



A10-7700K
CPU

FPU

MEM

[justapost]

Arauna

Benchmark	A8-6500	A10-7700K
Arauna (Cycles)	3068	3151	3%
Arauna			3%

A8-6500


A10-7700K

[justapost]

Cinebench

Benchmark	A8-6500	A10-7700K
8.5 single	451	461	2%
9.5 single	560	573	2%
10 single	3551	3713	5%
11.5 single	0.87	0.89	2%
15 single	79	79	0%
Cinebench single			2%
8.5 multi	1229	1371	12%
9.5 multi	1582	1686	7%
10 multi	11164	12139	9%
11.5 multi	2.86	3.14	10%
15 multi	263	273	4%
Cinebench multi			8%
8.5 OpenGL	6622	6834	3%
9.5 OpenGL	6944	7093	2%
10 OpenGL	7186	6818	-5%
11.5 OpenGL	33.0	38.83	18%
15 OpenGL	32.24	38.39	19%
Cinebench OpenGL			7%

8.5 / 9.5 / 10 / 11.5 / 15

A8-6500


A10-7700K

[justapost]

CompuBench

CPU Test

Benchmark	A8-6500	A10-7700K
Physics: SPH Fluid Simulation	636	707	11%
Physics			11%
Graphic: Raytrace	6215	6831	10%
Graphics			10%
Vision: Optical Flower	427	454	6%
Vision			6%
Image Filter: Sobel - Global Memory Usage	37226	38900	4%
Image Filter: Sobel - Local Memory Usage	15622	18607	19%
Image Filter: Sobel - Image Usage	4798	5381	12%
Image Filter: Separated Gaussian Blur - Global Memory Usage	8040	8188	2%
Image Filter: Separated Gaussian Blur - Image Usage	1683	1813	8%
Image Filter: Median 3x3 - Image Usage	2183	2319	6%
Image Filter			9%
Programming Principles: Bucketing - Parallel	813	993	22%
Programming Principles: Reduction- Parallel	4108	7161	74%
Programming Principles: Scanning- Parallel	1233	2043	66%
Programming Principles: Scanning- Sequential	43057	47253	10%
Programming Principles: Sum Global	456	379	-17%
Programming Principles: Sum Local	8028	8387	4%
Programming Principles: Bitonic Merge Sort	157	167	6%
Programming Principles:Tree Search	672	783	17%
Programming Principles			22%
Compubench CPU			12%

GPU Test

Benchmark	A8-6500	A10-7700K
Physics: SPH Fluid Simulation	1675	5053	202%
Physics			202%
Graphic: Raytrace	16366	51216	213%
Graphics			213%
Vision: Optical Flower	1450	3139	116%
Vision			116%
Image Filter: Sobel - Global Memory Usage	188100	382500	103%
Image Filter: Sobel - Local Memory Usage	75325	172700	129%
Image Filter: Sobel - Image Usage	286700	429799	50%
Image Filter: Separated Gaussian Blur - Global Memory Usage	23418	58100	148%
Image Filter: Separated Gaussian Blur - Image Usage	113300	158900	40%
Image Filter: Median 3x3 - Image Usage	107646	168200	56%
Image Filter			87%
Programming Principles: Bucketing - Parallel	1238	6148	397%
Programming Principles: Reduction- Parallel	39080	229700	488%
Programming Principles: Scanning- Parallel	13625	62400	358%
Programming Principles: Scanning- Sequential	938	9727	937%
Programming Principles: Sum Global	2104	28472	1253%
Programming Principles: Sum Local	52474	92000	75%
Programming Principles: Bitonic Merge Sort	1295	2548	96%
Programming Principles:Tree Search	648	1328	105%
Programming Principles			464%
Compubench CPU			216%

A8-6500


A10-7700K

[justapost]

CPUMark

Benchmark	A8-6500	A10-7700K
CPU Mark	394	413	5%
CPU Mark			5%

A8-6500


A10-7700K

[justapost]

Crysis Timedemo

Benchmark	A8-6500	A10-7700K
1280x720 low	8587.5	8117.5	-5%
1400x960 low	8380	7956	-5%
1920x1080 low	6642.5	6766	2%
1280x720 medium	5589	5408.5	-3%
1400x960 medium	4736	4100.5	-14%
1920x1080 medium	3409.5	3572	5%
1280x720 high	3337	3279	-2%
1400x960 high	2573.5	3045.5	18%
1920x1080 high	1769	2168.5	23%
1280x720 high 4AA	2705.5	2947.5	9%
1400x960 high 4AA	1992.5	2410.5	21%
1920x1080 high 4AA	1331	1670	25%
1280x720 veryhigh	1968.5	2347	19%
1400x960 veryhigh	1448	1797	24%
1920x1080 veryhigh	975	1230	26%
1280x720 veryhigh 4AA	1487	1768	19%
1400x960 veryhigh 4AA	1057.5	1310.5	24%
1920x1080 veryhigh 4AA	712.5	890.5	25%
Crysis			12%

A8-6500


A10-7700K

[justapost]

FritzChess

Benchmark	A8-6500	A10-7700K
Fritzchess 1 Core	1920	1948	1%
Fritzchess 2 Cores	3828	3870	1%
Fritzchess 3 Cores	4866	5358	10%
Fritzchess 4 Cores	5895	6779	15%
Fritzchess			7%

A8-6500


A10-7700K

[justapost]

Frybench

Benchmark	A8-6500	A10-7700K
Frybench (sec)	809	809	0%
Frybench			0%

A8-6500


A10-7700K

[justapost]

Geekbench

Single

Benchmark	A8-6500	A10-7700K
AES	2372	2326	-2%
Twofish	2344	2427	4%
SHA1	1900	2538	36%
SHA2	2342	2989	28%
BZip2 Compression	1970	2194	11%
BZip2 Decompression	2234	2374	6%
JPEG Compression	2263	2396	6%
JPEG Decompression	2558	2765	8%
PNG Compression	1841	1984	8%
PNG Decompression	2012	2174	8%
Soble	1852	1978	7%
Lua	1876	2157	15%
Dijkstra	1910	1943	2%
Integer			10%
BlackScholes	2075	2214	7%
Mandelbrot	2227	2225	0%
Sharpen Filter	2093	2109	1%
Blur Filter	1725	2001	16%
SGEMM	1262	1626	29%
DGEMM	1169	1623	39%
SFFT	1708	2385	40%
DFFT	1439	2110	47%
N-Body	1346	1656	23%
Ray Trace	1919	2355	23%
Floating Point			22%
Stream Copy	1651	1577	-4%
Stream Scale	1613	1524	-6%
Stream Add	1611	1502	-7%
Stream Triad	1662	1552	-7%
Memory			-6%
Single			9%

Multi

Benchmark	A8-6500	A10-7700K
AES	5419	5471	1%
Twofish	8859	9662	10%
SHA1	7361	10105	37%
SHA2	8695	11854	36%
BZip2 Compression	6441	7962	24%
BZip2 Decompression	7579	8764	16%
JPEG Compression	7895	8918	13%
JPEG Decompression	7925	9810	24%
PNG Compression	6746	7552	12%
PNG Decompression	6532	7769	19%
Soble	5982	6782	13%
Lua	6256	8067	29%
Dijkstra	5486	5980	9%
Integer			19%
BlackScholes	7143	7577	6%
Mandelbrot	8195	8262	1%
Sharpen Filter	5786	6968	20%
Blur Filter	4731	6288	33%
SGEMM	4571	5341	17%
DGEMM	4360	5249	20%
SFFT	4699	6135	30%
DFFT	4470	6625	48%
N-Body	4912	6154	25%
Ray Trace	6330	8638	36%
Floating Point			24%
Stream Copy	2418	2396	-1%
Stream Scale	2402	2355	-2%
Stream Add	2472	2294	-7%
Stream Triad	2563	2373	-7%
Memory			-4%
Multi			13%

A8-6500
http://browser.primatelabs.com/geekbench3/362422

A10-7700K
http://browser.primatelabs.com/geekbench3/365673