SIX is Sexy! (Unboxed)

[richierich]

Wow ridney, awesome findings.

[MadDias]

whau flank3r! i'm running on water (see my sig), the problem with my mobo is that it can't disable turbo (yet) so i'm stuck

what bios are u running on ur gd70? the newest is 1.B1 or 1.11B1
maybe there is a setting for disabling turbo in the newest beta.

edit: ok... just saw in the first post that ur on 1.10B8
i guess the only thing they changed in 1.11b1 is c´n´q working with adjusted cpu multiplier on 955 and 965 cpus

[ReKcOlNu]

Turbo ran on FOUR cores!!!

According to AMD only 3 cores can be in turbo mode at a time, and since turbo is controlled by the CPU itself I doubt 4 cores are getting boosted, I'd say CPU-Z is a bad tool to use since it doesn't update often enough to see whats actually going on.

For example on a dual core Phenom II @ 3.4GHz (CnQ enabled) I ran 2 instances of CPU-Z (one for each core) and at the same time ran a program set to take 10 samples per second, the yellow line is P-state (in percentage) and the blue line is the MHz usage, watching core 0 CPU-Z only detected the CPU speed changing two times when it changed at least 10 times.



Thanks for sharing your X6 results!
.

[ridney]

Besides, how that Turbo thing scales?

turbo is scaling quite nicely between one to three cores fast, i'm pretty impressed with the hardware based turbo so far

ridney: its beter for OC than C3 Denebs, right? Hope, with refresh BIOS u will can turbo disabled for testing. But u are lucky man . What about wprime benchmarks?

i didn't had a c3 but a c2, so far i can say that it's nothing like it, it's a lot better in my opinion so i guess it is better than c3 (i'm on normal water cooling not chilled water)

what bios are u running on ur gd70? the newest is 1.B1 or 1.11B1

since friday i have been runing the latest 1.11b1 bios

According to AMD only 3 cores can be in turbo mode at a time, and since turbo is controlled by the CPU itself I doubt 4 cores are getting boosted, I'd say CPU-Z is a bad tool to use since it doesn't update often enough to see whats actually going on.

maybe someone can post a phenom II quad at 4.4ghz cpu score of 3dmark06 so we can compare the score i got and formulate an opinion based from those results



oh and i have some interesting bit found just this morning. turned on pc and checked cpu-z then saw i was only getting x11 multi on a 270 bus, so i was kinda shocked and worried that something bad might have happened (crashed system before going to bed), this is the first time i saw an x11 multi so i ran pi to check it again and took a screenie. it really is getting x11 on core#4



maybe a buggy bios? idk

[richierich]

weird.

[FlanK3r]

rideny: i can post only D06 at 4.2 GHz , its max running for my aircooling...

[ridney]

flank3r, i'm sure 4.2 on your mobo is pretty easy. temps are not so much of a problem, thubans are really cool!

okay, i tried pov-ray render test but i'm not sure if i'm doing this correctly



EDIT: updated with x264 bench result

x264 HD BENCHMARK 3.0 RESULTS 
 
Please do not compare it with older versions of the benchmark! 
Please copy/paste everything below the line to to report your data 
 
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 
 
Results for x264.exe r1342 
-------------------------- 
encoded 1442 frames, 87.62 fps, 3898.70 kb/s
encoded 1442 frames, 87.37 fps, 3898.70 kb/s
encoded 1442 frames, 87.45 fps, 3898.70 kb/s
encoded 1442 frames, 87.20 fps, 3898.70 kb/s
encoded 1442 frames, 34.95 fps, 3971.69 kb/s
 
 
System Details 
-------------- 
	Name			AMD K10
	Codename		
	Specification		AMD Phenom(tm) II X6 1055T Processor
	Specification		PC3-12800
	Specification		PC3-12800
	Core Stepping		
	Stock frequency		3300 MHz
	Core Speed		3710.2 MHz
	FID range		5.0x - 16.5x
 
Northbridge			AMD 790FX rev. 00
Southbridge			ATI SB750 rev. 00
 
CAS# latency (CL)		7.0
RAS# to CAS# delay (tRCD)	7
RAS# Precharge (tRP)		7
Cycle Time (tRAS)		24
Command Rate (CR)		1T
	Command Rate		2T
	Command Rate		2T
Memory Frequency		706.7 MHz (3:8)
Memory Type			DDR3
Memory Size			4096 MBytes
Channels			Dual, (Unganged)
 
Windows Version			Microsoft Windows 7 (6.1) Ultimate Edition   (Build 7600) 
 
Number of processors		1
Number of threads		6
	Number of threads	6 (max 6)
	L2 cache		6 x 512 KBytes, 16-way set associative, 64-byte line size
	Instructions sets	MMX (+), 3DNow! (+), SSE, SSE2, SSE3, SSE4A, x86-64
	Package 		Socket AM2+ (940)
	Package Type		0x1
 
	Temperature 0	33°C (91°F) [0x21] (TMPIN0)
	Temperature 1	54°C (129°F) [0x36] (TMPIN1)
	Temperature 2	46°C (114°F) [0x2E] (TMPIN2)
	Temperature 0	31°C (86°F) [0xF4] (Core #0)
	Temperature 1	31°C (87°F) [0xF5] (Core #1)
	Temperature 2	31°C (87°F) [0xF6] (Core #2)
	Temperature 3	31°C (88°F) [0xF9] (Core #3)
	Temperature 4	31°C (88°F) [0xF9] (Core #4)
	Temperature 5	31°C (88°F) [0xFA] (Core #5)

[crazydiamond]

w7 64bit 4 cores @ 4335mhz, nb @ 2805mhz, ram a lil under 1360mhz c8 got me cpu score 6038

[lv_88]

I would've loved to see a cinebench r10 and 11.5 results at 4Ghz if thats possible, or maybe its already been done?
looks like its a fun cpu this far!

[FlanK3r]

this wwas my run a few days ago (aircooled, not max CPU, only maybe max with this voltage) CPU= 6049

[ridney]

nice score flank3r! so cpu score = 6409 at 4.1ghz and mine's 6808 at 4.45ghz (4 cores) is plausible i don't think it would have gotten that high if it was running turbo on 3 cores though

okay, here's a cinebench 11.5 run. booted at 3.5ghz, oc dialed 4 clicks up at 5mhz HTT/click to 3.7 before i started the run. during the start when turbo went off and loaded all six cores at 3.7, i oc dialed again 3 clicks to get up to 3.99 and then before the end i dialed back again 3 clicks down (turbo kicks in at the end of the run and would have crashed, 4.7ghz turbo )

[richierich]

nice results ridney, and gotta love your way of adjusting the htt before turbo kicks in at the end, hahaha.

[Lightman]

What volts for that 3.99GHz run Ridney?
I don't believe CPU-Z reading

I must have .... now

[FlanK3r]

rideny: thank u, i have some better CPU score...Its older result

[AMD-C]

According to AMD only 3 cores can be in turbo mode at a time, and since turbo is controlled by the CPU itself I doubt 4 cores are getting boosted, I'd say CPU-Z is a bad tool to use since it doesn't update often enough to see whats actually going on.

For example on a dual core Phenom II @ 3.4GHz (CnQ enabled) I ran 2 instances of CPU-Z (one for each core) and at the same time ran a program set to take 10 samples per second, the yellow line is P-state (in percentage) and the blue line is the MHz usage, watching core 0 CPU-Z only detected the CPU speed changing two times when it changed at least 10 times.



Thanks for sharing your X6 results!
.

It should be easy enough to figure out what's really going on here directly via bench results.

We just need the X6 to run Cinebench limited to 4 threads instead of the default. Since Cinebench scales so preditcably with the total clock available to it, the percentage of score difference should be 4%, the same as the percentage of clock difference between 3 turbo cores + 1 stock core versus 4 turbo cores.

Anyone with an X4 can simulate this with AOD by setting one core down by 2.5 multipliers.

[AMD-C]

flank3r, i'm sure 4.2 on your mobo is pretty easy. temps are not so much of a problem, thubans are really cool!

FYI, for reference.

My X4 at 4.1Ghz gives the following:

POV-RAY 3.7 - 3997.13 PPS (1019.41 PPS CPU time)

x264 HD BENCHMARK 3.0 RESULTS 
 
Please do not compare it with older versions of the benchmark! 
Please copy/paste everything below the line to to report your data 
 
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 
 
Results for x264.exe r1342 
-------------------------- 
encoded 1442 frames, 87.32 fps, 3900.68 kb/s
encoded 1442 frames, 87.84 fps, 3900.69 kb/s
encoded 1442 frames, 87.11 fps, 3900.68 kb/s
encoded 1442 frames, 88.24 fps, 3900.68 kb/s
encoded 1442 frames, 26.72 fps, 3971.31 kb/s
encoded 1442 frames, 26.78 fps, 3970.96 kb/s
encoded 1442 frames, 26.75 fps, 3971.90 kb/s
encoded 1442 frames, 26.73 fps, 3971.53 kb/s
 
 
System Details 
-------------- 
	Name			AMD Phenom II X4 965
	product			System Product Name
	Codename		Deneb
	Specification		AMD Phenom(tm) II X4 965 Processor
	Core Stepping		RB-C3
	Technology		45 nm
	Core Speed		4100.0 MHz
 
Northbridge			AMD 790X rev. 00
Southbridge			ATI SB750 rev. 00
 
CAS# latency (CL)		8.0
RAS# to CAS# delay (tRCD)	8
RAS# Precharge (tRP)		8
Cycle Time (tRAS)		20
Command Rate (CR)		1T
Memory Frequency		800.0 MHz (1:4)
Memory Type			DDR3
Memory Size			4096 MBytes
Channels			Dual, (Unganged)
 
Windows Version			Microsoft Windows 7 (6.1) Enterprise Edition  (Build 7600) 
 
Number of processors		1
Number of threads		4
	Number of threads	4 (max 4)
	L2 cache		4 x 512 KBytes, 16-way set associative, 64-byte line size
	Instructions sets	MMX (+), 3DNow! (+), SSE, SSE2, SSE3, SSE4A, x86-64
	Package 		Socket AM2+ (940)
	Package Type		0x1
 
	Temperature 0	39°C (102°F) [0x27] (TMPIN0)
	Temperature 1	24°C (75°F) [0x18] (TMPIN1)
	Temperature 0	47°C (117°F) [0x17B] (Core #0)
	Temperature 1	48°C (117°F) [0x17C] (Core #1)
	Temperature 2	48°C (117°F) [0x17D] (Core #2)
	Temperature 3	48°C (118°F) [0x17F] (Core #3)

The fact that our first pass results of x264 makes me wonder if it's being limited to 4 threads? Your X6 running 4 threads at 3x4372 + 1x 3710 would only have 2% more total clock than 4x4100, but it looks like your second pass it much faster, and at 35fps is exactly what one would expect from the 6x3710.

Strange.

[ridney]

i think i messed up (ran with errors on the second pass) with the ram settings when i did the pov-ray bench, so here's another result at 270HTT, 3.8 all cores, 4.4 turbo

x264 HD BENCHMARK 3.0 RESULTS 
 
Please do not compare it with older versions of the benchmark! 
Please copy/paste everything below the line to to report your data 
 
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 
 
Results for x264.exe r1342 
-------------------------- 
encoded 1442 frames, 88.97 fps, 3898.70 kb/s
encoded 1442 frames, 89.57 fps, 3898.70 kb/s
encoded 1442 frames, 89.57 fps, 3898.70 kb/s
encoded 1442 frames, 89.39 fps, 3898.70 kb/s
encoded 1442 frames, 35.91 fps, 3968.57 kb/s
encoded 1442 frames, 36.00 fps, 3971.60 kb/s
encoded 1442 frames, 35.84 fps, 3960.24 kb/s
encoded 1442 frames, 35.87 fps, 3954.02 kb/s
 
 
System Details 
-------------- 
	Name			AMD K10
	Codename		
	Specification		AMD Phenom(tm) II X6 1055T Processor
	Specification		PC3-12800
	Specification		PC3-12800
	Core Stepping		
	Stock frequency		3300 MHz
	Core Speed		4455.2 MHz
	FID range		5.0x - 16.5x
 
Northbridge			AMD 790FX rev. 00
Southbridge			ATI SB750 rev. 00
 
CAS# latency (CL)		7.0
RAS# to CAS# delay (tRCD)	7
RAS# Precharge (tRP)		7
Cycle Time (tRAS)		24
Command Rate (CR)		1T
	Command Rate		2T
	Command Rate		2T
Memory Frequency		720.0 MHz (3:8)
Memory Type			DDR3
Memory Size			4096 MBytes
Channels			Dual, (Unganged)
 
Windows Version			Microsoft Windows 7 (6.1) Ultimate Edition   (Build 7600) 
 
Number of processors		1
Number of threads		6
	Number of threads	6 (max 6)
	L2 cache		6 x 512 KBytes, 16-way set associative, 64-byte line size
	Instructions sets	MMX (+), 3DNow! (+), SSE, SSE2, SSE3, SSE4A, x86-64
	Package 		Socket AM2+ (940)
	Package Type		0x1
 
	Temperature 0	29°C (84°F) [0x1D] (TMPIN0)
	Temperature 1	50°C (121°F) [0x32] (TMPIN1)
	Temperature 2	42°C (107°F) [0x2A] (TMPIN2)
	Temperature 0	27°C (80°F) [0xD7] (Core #0)
	Temperature 1	27°C (81°F) [0xDA] (Core #1)
	Temperature 2	28°C (81°F) [0xDD] (Core #2)
	Temperature 3	28°C (81°F) [0xDE] (Core #3)
	Temperature 4	28°C (82°F) [0xE1] (Core #4)
	Temperature 5	29°C (83°F) [0xE4] (Core #5)

here's another one for pov-ray




downloaded spec_3dsmax and spec_viewperf and not sure to run this thing

[ridney]

What volts for that 3.99GHz run Ridney?
I don't believe CPU-Z reading

I must have .... now

1.525v at bios, i think 1.475v is okay for 4.0Ghz but i was just making sure

FYI, for reference.

The fact that our first pass results of x264 makes me wonder if it's being limited to 4 threads? Your X6 running 4 threads at 3x4372 + 1x 3710 would only have 2% more total clock than 4x4100, but it looks like your second pass it much faster, and at 35fps is exactly what one would expect from the 6x3710.

Strange.

I guess first pass runs on turbo and second pass runs on all cores, that should explain the similarity on the first pass scores with your quad.

[crazydiamond]

i think black editions will be money with these cpus because as long as turbo can stay on with changed multi, we should be able to get some nice x6 clocks and still use the 400mhz turbo. when the htt starts getting that high and turbo gets to 600+mhz i think it is easier for the system to become unstable. something like 4.0 with 4.4 turbo, around 1.5v like c3 denebs need

[AMD-C]

i think black editions will be money with these cpus because as long as turbo can stay on with changed multi, we should be able to get some nice x6 clocks and still use the 400mhz turbo. when the htt starts getting that high and turbo gets to 600+mhz i think it is easier for the system to become unstable. something like 4.0 with 4.4 turbo, around 1.5v like c3 denebs need

Yes.

Ideally I'd like to see BEs sport a configurable Turbo multiplier. That would be a truly killer feature.

I'm tempted to pre-order a 1090T right now, but I think I'm going to wait and see if they will pop up at Microcenter on release day. If so, I'll just pick one up.

[[XC] Oj101]

I'm officially excited but beginning to wonder of six cores are enough for me

[Lightman]

I'm officially excited but beginning to wonder of six cores are enough for me

Good thinking on your part

[[XC] Oj101]

Sorry, typo:

I'm officially excited but beginning to wonder if six cores are enough for me

[ridney]

nothing is enough, we always want moar

[ridney]

i accidentally updated my mobo with an earlier bios (a7577ams.190 dated around feb 2010) and was quite surprised that it still POSTed. Got into bios and checked, processor was unknown but the interesting part was that the cpu multi was up to x16.5 instead of x14. Exciting! So i tried the highest multi (x16.5) and booted to windows only to find out that it was stuck at x4 (800mhz) :P: No matter what i did, it was always running at x4 grrr i thought i outsmarted the cpu and turned off turbo

anyways, here are some more bench results

winrar




7-zip




if you guys find something about turning turbo off, i appreciate if you can post some info here