My 8150 Bulldozer experience - so far!

[soccerballtux]

But the Thuban x6 always ties or bests the 8150 does it not? Doesn't that make it horrible?

 

[AtenRa]

But the Thuban x6 always ties or bests the 8150 does it not? Doesn't that make it horrible?

FX 8150 is faster than Thuban x6 in MT with less idle power, at load they have the same performance/watt.

http://www.anandtech.com/show/4955/the-bulldozer-review-amd-fx8150-tested/7

 

[Yuriman]

Which isn't bad, but isn't great either considering Bulldozer is 32nm and Thuban is 45nm.

 

[guskline]

My 1100 Thuban was fast but the 8150 OCs much more and is overall faster

 

[guskline]

Changed the configuration to 21 multiplier, 215 fsb and 1.4 v for the cpu. Runs stable and extremely cool at 4.5 Ghz.

 

[Makaveli]

Changed the configuration to 21 multiplier, 215 fsb and 1.4 v for the cpu. Runs stabler and extremely cool at 4.5 Ghz.

Is stabler a word ?

And no such thing as an extremely cool bulldozer ()

Glad to see you are enjoying your build and the thread has stayed on topic the last couple pages !!

 

[Durvelle27]

@Guskline i'll trade cpu and motherboards with you XD

 

[guskline]

Is stabler a word ?

And no such thing as an extremely cool bulldozer ()

Glad to see you are enjoying your build and the thread has stayed on topic the last couple pages !!

Sorry for the "stabler". I meant stable.
Sensor Properties: Sensor Type ITE IT8721F (ISA 290h) GPU Sensor Type Diode, Volterra VT1165 (ATI-Diode, 70h, 70h) Motherboard Name Asus Sabertooth 990FX Chassis Intrusion Detected No Temperatures: Motherboard 31 °C (88 °F) CPU 32 °C (90 °F) CPU #1 / Core #1 21 °C (70 °F) CPU #1 / Core #2 21 °C (70 °F) CPU #1 / Core #3 21 °C (70 °F) CPU #1 / Core #4 21 °C (70 °F) CPU #1 / Core #5 21 °C (70 °F) CPU #1 / Core #6 21 °C (70 °F) CPU #1 / Core #7 21 °C (70 °F) CPU #1 / Core #8 21 °C (70 °F) North Bridge 38 °C (100 °F) DIMM 32 °C (90 °F) PCI-E #1 36 °C (97 °F) PCI-E #2 33 °C (91 °F) USB 3.0 #1 35 °C (95 °F) USB 3.0 #2 34 °C (93 °F) GPU1: GPU Diode (DispIO) 38 °C (100 °F) GPU1: GPU Diode (MemIO) 36 °C (97 °F) GPU1: GPU Diode (Shader) 38 °C (100 °F) GPU1: GPU VRM 37 °C (99 °F) GPU2: GPU Diode (DispIO) 40 °C (104 °F) GPU2: GPU Diode (MemIO) 47 °C (117 °F) GPU2: GPU Diode (Shader) 41 °C (106 °F) GPU2: GPU VRM 35 °C (95 °F) Temperature #1 43 °C (109 °F) Temperature #2 42 °C (108 °F) WDC WD5000AAKX-003CA0 33 °C (91 °F) Cooling Fans: CPU 2163 RPM GPU1 2373 RPM (35%) GPU2 2149 RPM (35%) Voltage Values: CPU Core 1.392 V CPU VDDA 2.487 V +3.3 V 3.336 V +5 V 5.038 V +12 V 11.952 V VDDA 2.532 V North Bridge Core 1.100 V HyperTransport 1.197 V DIMM 1.519 V GPU1: GPU Core 0.950 V GPU1: GPU VRM 0.950 V GPU2: GPU Core 1.000 V GPU2: GPU VRM 1.000 V Current Values: GPU1: GPU VRM 1.74 A GPU2: GPU VRM 1.74 A
This is info on my machine at 4.5Ghz.

 

[guskline]

@Guskline i'll trade cpu and motherboards with you XD

Afraid not!

 

[Durvelle27]

Afraid not!

D:

 

[guskline]

Makaveli: Perhaps the word Extremely with cool was pushing it, but considering the the perception that an OC'd BD runs constantly like a blast furnace, I wanted to show that with good cooling it runs cool at idle. Obviously it heats up when I run, for example, Intel Burn Test ( more stress than Prime95). I keep reading and tweeking to find the coolest running combination that is OC'd to 4.5 Ghz. When I go to 4.6 Ghz or above, power usage and heat buildup increase dramatically. I'm really intersted in seeing what tricks, if any, AMD employed in PileDriver. The reading I've done on Resonant mesh technology appears to indicate that the CPU runs cooler but there is a limited Mhz range that it works so higher OCing might not be in the cards. We shall see.

 

[SlowSpyder]

Sorry for the "stabler". I meant stable.
Sensor Properties: Sensor Type ITE IT8721F (ISA 290h) GPU Sensor Type Diode, Volterra VT1165 (ATI-Diode, 70h, 70h) Motherboard Name Asus Sabertooth 990FX Chassis Intrusion Detected No Temperatures: Motherboard 31 °C (88 °F) CPU 32 °C (90 °F) CPU #1 / Core #1 21 °C (70 °F) CPU #1 / Core #2 21 °C (70 °F) CPU #1 / Core #3 21 °C (70 °F) CPU #1 / Core #4 21 °C (70 °F) CPU #1 / Core #5 21 °C (70 °F) CPU #1 / Core #6 21 °C (70 °F) CPU #1 / Core #7 21 °C (70 °F) CPU #1 / Core #8 21 °C (70 °F) North Bridge 38 °C (100 °F) DIMM 32 °C (90 °F) PCI-E #1 36 °C (97 °F) PCI-E #2 33 °C (91 °F) USB 3.0 #1 35 °C (95 °F) USB 3.0 #2 34 °C (93 °F) GPU1: GPU Diode (DispIO) 38 °C (100 °F) GPU1: GPU Diode (MemIO) 36 °C (97 °F) GPU1: GPU Diode (Shader) 38 °C (100 °F) GPU1: GPU VRM 37 °C (99 °F) GPU2: GPU Diode (DispIO) 40 °C (104 °F) GPU2: GPU Diode (MemIO) 47 °C (117 °F) GPU2: GPU Diode (Shader) 41 °C (106 °F) GPU2: GPU VRM 35 °C (95 °F) Temperature #1 43 °C (109 °F) Temperature #2 42 °C (108 °F) WDC WD5000AAKX-003CA0 33 °C (91 °F) Cooling Fans: CPU 2163 RPM GPU1 2373 RPM (35%) GPU2 2149 RPM (35%) Voltage Values: CPU Core 1.392 V CPU VDDA 2.487 V +3.3 V 3.336 V +5 V 5.038 V +12 V 11.952 V VDDA 2.532 V North Bridge Core 1.100 V HyperTransport 1.197 V DIMM 1.519 V GPU1: GPU Core 0.950 V GPU1: GPU VRM 0.950 V GPU2: GPU Core 1.000 V GPU2: GPU VRM 1.000 V Current Values: GPU1: GPU VRM 1.74 A GPU2: GPU VRM 1.74 A
This is info on my machine at 4.5Ghz.

Any power use numbers at 4.5GHz? My Thuban is pretty power hungry at 4GHz/2.6Ghz.

 

[Makaveli]

Makaveli: Perhaps the word Extremely with cool was pushing it, but considering the the perception that an OC'd BD runs constantly like a blast furnace, I wanted to show that with good cooling it runs cool at idle. Obviously it heats up when I run, for example, Intel Burn Test ( more stress than Prime95). I keep reading and tweeking to find the coolest running combination that is OC'd to 4.5 Ghz. When I go to 4.6 Ghz or above, power usage and heat buildup increase dramatically. I'm really intersted in seeing what tricks, if any, AMD employed in PileDriver. The reading I've done on Resonant mesh technology appears to indicate that the CPU runs cooler but there is a limited Mhz range that it works so higher OCing might not be in the cards. We shall see.

Your temps look good, also just a tip trying to put some spacing in the temp data it just looks like a giant wall of text. Would make it easier to read for others.

I had a friend that was running one of those Dual core P4 back in the day and damn did that heat up a room. AMD has been good with the idle temps its just the full load on BD that it gets crazy. I think the next spin of this arch will reduce that alot as its the other sore thumb sticking out besides single threaded performance.

I'm not sure if it was asked but what is your max overclock at stock vcore and what are your temps like?

 

[guskline]

Makaveli: First sorry for the run on text. I was trying to figure out how to copy a screen from AIDA 64 and the run on text was the best I could do (perhaps someone running that program could help me learn how to paste the results in a more readable fashion.)

Max OC at stock, HMMM? I originally had a CM Hyper 212+ with Push/Pull in an old Chieftech tower with 2 80mm exhaust fans an 1 80 mm intake. Temps at stock settings were fine and I think the 4.2 GHz on all 8 cores was stable at 1.4 v with temps not exceeding 65C. Honestly I never tried seeing the max OC at stock voltage.

I have decided to try lowering the OC voltage to 1.4 v and tweak setting to get the best speed without excessive heat. Right now I'm running 21 multiplier by 215 fsb 4.515Ghz with LLC set to extreme, cpu voltage at 1.4v, CPU-NB and HT at 2356?(I think) but temps, even running Intel Burn Test and Prime 95, not exceeding 65C. The vcores heat up but so far no crashes etc and stability is rock solid. BTW, the 65C is the MAX overall cpu reading, the individual cores read @55-57 C max.

I have the H100 set for Max performance drawing cool air in with 1 Antec 120 mm fan exhaust at 2000rpm and 2 120 mm intakes at @1500 rpms in a CM 912HAF case.
I "altered" the case by cutting out part of the top grill to allow the radiator part of the Corsair H100 to mount on topof the case with the 2 fans provided on top of it drawing cool air into the radiator and having the Antec 120mm fan pull out the hot air in the rear exhaust.

I'm ordering a better Panaflow rear exhaust 120 mm fan (max 2700rpm) and a single CM 200 mm intake fan for the front. these fans are controlled by a fan controller. I've found that cranking the 8150 beyond @4.4 Ghz requires much more attention to cooling. I'm sure others have gone higher with acceptable temps but I'd bet the house that they used improved cooling techniques.

Suffice it to say that this 8150 build is truly a "work in progress" trying to find the right balance between OCing and heat. Heck I could have just kept the chip at stock with a stock cooler, but what fun would that be?

 

[SlowSpyder]

What does the average 8150 reach? I've seen reviews at 4.7-4.8GHz, but I'm not sure if that's realistic for most 8150's. I have never thought Bulldozer's performance was really that bad, but the power needed for those overclocks seems to get out of hand quickly. I know my Thuban does the same thing.

 

[guskline]

What does the average 8150 reach? I've seen reviews at 4.7-4.8GHz, but I'm not sure if that's realistic for most 8150's. I have never thought Bulldozer's performance was really that bad, but the power needed for those overclocks seems to get out of hand quickly. I know my Thuban does the same thing.

SlowSpyder, good point. What does the average 8150 reach with stock voltage at safe temps? I SERIOUSLY doubt in the 4.7-4.8 range. I would venture to say that SAFELY you won't get more than 4.2 GHZ for ALL 8 cores on stock cooling at stock voltage. Remember the chip runs 3.6 on 8 cores, 3.9 Turbo (presumably 8 cores) but 4.2 Ghz only on a MAX of 4(perhaps 3) cores to keep the heat and voltage within specs. When I refer to my 4.515 Ghz speed that is ALL 8 cores. That equates to more heat produced and more power needed. To run 4.7 to 4.8 Ghz on all 8 cores, based on what I've experienced you better have the cream of the crop 8150 with an advanced water cooling system, best MB, solid ram and tons of ventilation. My chip seems to run safely, and stable at 4.5 Ghz. I can run fine at 4.6 Ghz but when I run Intel Burn Test the temps are too high. Yep it says the system is solid and safely OC'd but I don't like the temps. I'd rather sacrifice a .1Ghz bump and know I'm solid for 2 hrs of solid game play than try to tweak too much and get a BSOD just for bragging rights. Perhaps other 8150 owners can chime in.

 

[SlowSpyder]

SlowSpyder, good point. What does the average 8150 reach with stock voltage at safe temps? I SERIOUSLY doubt in the 4.7-4.8 range. I would venture to say that SAFELY you won't get more than 4.2 GHZ for ALL 8 cores on stock cooling at stock voltage. Remember the chip runs 3.6 on 8 cores, 3.9 Turbo (presumably 8 cores) but 4.2 Ghz only on a MAX of 4(perhaps 3) cores to keep the heat and voltage within specs. When I refer to my 4.515 Ghz speed that is ALL 8 cores. That equates to more heat produced and more power needed. To run 4.7 to 4.8 Ghz on all 8 cores, based on what I've experienced you better have the cream of the crop 8150 with an advanced water cooling system, best MB, solid ram and tons of ventilation. My chip seems to run safely, and stable at 4.5 Ghz. I can run fine at 4.6 Ghz but when I run Intel Burn Test the temps are too high. Yep it says the system is solid and safely OC'd but I don't like the temps. I'd rather sacrifice a .1Ghz bump and know I'm solid for 2 hrs of solid game play than try to tweak too much and get a BSOD just for bragging rights. Perhaps other 8150 owners can chime in.

I'd do the same thing. You won't notice 100MHz, except very slightly better scores in benchmarks. But in real world use, you'd never notice a difference. My Thuban can hit 4.2GHz on 5 of the six cores, I could try more voltage, but I figure there's not much point to it.

 

[guskline]

I'd do the same thing. You won't notice 100MHz, except very slightly better scores in benchmarks. But in real world use, you'd never notice a difference. My Thuban can hit 4.2GHz on 5 of the six cores, I could try more voltage, but I figure there's not much point to it.

Smart move!:sneaky: Besides with your 7970 GPU you have PLENTY of power.

 

[guskline]

Made some fan changes and also backed off the clock slightly. I replaced 2 120 mm fans with a single CM 200 mm fan in the front which pushes a lot of air but quieter than 2 120 mms. in the rear I was going to replace the 2000 rpm antec 120mm fan with a 2700rpm 120mm panaflow but it is too noisy and draws too much power ( I only have an Antec TP II 750W). I also decided to back off to 4.4Ghz (21 x 210) to run a little cooler at max. The settings I mofified in BIOS are CPU LLC=UltraHigh; CPU/NB LLC=High; CPU Current Capacity=130%; CPU/NB Current Capacity=130%; CPU Phase control=Extreme(100%); CPU Voltage=fixed 1.393750; NB V=1.125 and Dram V = 1.51875,

The system runs cooler at max and I can keep the fan speed at a slightly lower speed.
Ironically I found running at 21x215 was just at cool so back to 4.5Ghz!

 

[jhu]

Is stabler a word ?

It's a perfectly cromulent word. It embiggens all of us.

 

[guskline]

cromulent
(Adj.) Used to describe a dubious or made up word, term, or phrase that is entirely plausible because it makes logical sense within existing language conventions.

Ironically, the word comulent itself is not cromulent. (See incromulent)

Origin: Television writer David X. Cohen for The Simpsons episode 3F13, "Lisa the Iconoclast".
My friend didn't believe me when I told her that meese was the plural of moose, but she admitted that it was entirely cromulent.



this term originated in the Animated Television Series "The Simpsons"
Edna: Embiggens? I never heard that word before I moved to Springfield.

Ms.Hoover: I don't know why. It's a perfectly cromulent word.

Jhu: I love it. I knew I should have watched the Simpsons more than I did!

 

[thilanliyan]

I really don't like the power consumption of my 6100 either. I have had to back off to 4.2GHz from 4.4GHz during the summer to keep temps in check. I don't notice much of a difference though so I will likely keep it like that. I hope Piledriver is a good improvement in the power consumption department.

 

[guskline]

I decided to "retire" the 2 5850s. My thee monitor setup has 2 identical EVGA GTX670 FTWs in SLI and my other rig has a PNY GTX680. I decided to run the 3 monitor setup with the single GTX680 (still plenty of power for 3 1920x1080 monitors) and use each of the GTX670 FTWs in the other 2500k rig and the Bulldozer rig. My 3DMark11 score jumped a 1000 points in performance from 7300P for the 5850s in CF to 8300 for the single GTX670FTW. The GPUs are NOT OC'd. Funny thing is the identical card in the 2500k rig also clocked to 4.5GHZ scores 9012P vs 8325 for the Bulldozer at 4.5Ghz. The ssd in the intel rig is faster (Intel 520) vs Intel X25M-80 in the Bulldozer rig. Suffice it to say the Intel sysem scores higher.

The nice thing is I am able to really compare the single monitor 8150 rig vs the 2500k rig with the same video card. The Intel rig shows faster scores but with identical high end video cards the Bulldozer seems VERY close in game play. Plus the single GTX 670 uses less power than the 2 5850s. Now I'm going to sell the 5850s!

 

[infoiltrator]

Not many more changes possible.
Go to a newer SSD maybe?

 

[guskline]

Not many more changes possible.
Go to a newer SSD maybe?

Thinking of that if both 5850s sell.