SuperPI 32M for testing CPU and system speed and stability

[Eldonko]

Great job hitechjb1! I looks forward to seeing more results so my Diego is not the only one posted hehe. Subscribed to yet another of your threads

 

[Sucka]

This really shouldn't be posted in the AMD section, as several of the best scores are Intels already. Intel may not do the best at 1m, but they really shine at high freq and 32m. I think i'll take my FX on air out for a whirl, it does 2.8 on air, should be good for a nice score i would think. I'll ask my Venice to play, and my San Diego when it gets here. I love these kinds of threads! Great work on the updating too. This thread needs to go to General CPU and get sticky action!

 

[Sucka]

CPU: ClawHammer FX-55 CAA2C 0439
memory: G. Skill 4400 LE 2 x 512 MB (Samsung TCCD)
motherboard: DFI LanParty UT NForce4 SLI
cooling: Zalman 7000cu
OS: Windows XP Professional SP1

- CPU: 2.80 GHz = 280 MHz x 10, 1.6 V
- memory: 280 MHz, 2.5-3-3-6 1T, 2.7 V
- SuperPI 32M completed in 26 min 14 sec

 

[El<(')>Maxi]

CPU: 4000 San Diego (sorry no chip details, 0515 week I think)
Memory: Mushkin LII 2 x 512 MB BH-5
Motherboard: DFI LanParty UT NForce4 Ultra-D BIOS 4.14.3
Cooling: Baker SS loaded @ -58°
OS: Windows Server 2003

- CPU: 3.18 GHz = 265 MHz x 12, 1.61 V will complete all PI runs @ 1.57v
- Memory: 265 MHz, 1.5-2-2-3 1T, 3.73 V
- SuperPI 32M completed in 23 min 29 Sec



*more voltage not helping at this point, first night of real testing on this new chip.

 

[HousERaT]

Here is one of my first.... not too shabby.

CPU: A64 3000+ Venice LBBLE 0516 CPBW
Memory: TwinMOS Speed Premium PC3200 UTT
Motherboard: DFI LanParty UT NForce4 Ultra-D BIOS 3/10p
Cooling: XP-90/Enermax 92mm
OS: Windows XP Pro SP2

CPU: 2.7 GHz = 245 x 11 (with Divider)
Memory: 245 MHz, 2-2-2-6 1T, 3.2v
SuperPI 32M completed in 26 min 54 Sec

http://www.ocforums.com/attachment.php?attachmentid=41496&stc=1

 

[g0dM@n]

You mean 300x9 with a divider? What vcore, Housrat?

 

[rdrash]

ADA4000DAA5BN, CABHE, 0515SPMW (1235306D50216)
G.skill Pc4400 LE (2x512) 250Mhz, 2.5 tCAS, 3 tRCD, 3tRP, 7tRAS, 1T cmd_rate, 2.6volts
DFI Lanparty NF4 Ultra-D, bios 4/14-3
Stock HSF
Windows Xp Professional x64 Edition Build 3790(service pack 1, v.1289)
Super Pi 32M completed in 26 minutes 59 seconds

Screen shots

 

[HousERaT]

You mean 300x9 with a divider? What vcore, Housrat?

right, 300x9 with a divider..... vcore shows 1.68v in bios and 1.66v in windows.

 

[Eldonko]

Bump. Lets see some more 32M!

 

[Molester]

well, i should update

this is my best
http://putfile.com/pic.php?pic=3/8316085617.jpg&s=x2

CPU: Athlon XP-M 2500+ IQYHA 0343 MPMW
memory: Corsair 3500 2 x 256 MB (BH-5)
motherboard: DFI LanParty NForce2 Ultra-B
cooling: Water, D-Tek White Water, Black Ice Extreme, Single-bay res, all 1/2"ID
OS: Windows XP Home SP2

- CPU: 2.62 GHz = 250 MHz x 10.5, 1.900 V vcore
- memory: 250 MHz, 2.0-2-2-11 1T, 3.3 V vdimm, 1.8 V vdd
- SuperPI 32M completed in 33 min 25 sec


this is stable
http://putfile.com/pic.php?pic=5/12722010073.jpg&s=x2

CPU: Athlon XP-M 2500+ IQYHA 0343 MPMW
memory: Corsair 3500 2 x 256 MB (BH-5)
motherboard: DFI LanParty NForce2 Ultra-B
cooling: Water, D-Tek White Water, Black Ice Extreme, Single-bay res, all 1/2"ID
OS: Windows XP Home SP2

- CPU: 2.60 GHz = 236 MHz x 11, 1.900 V vcore
- memory: 236 MHz, 2.0-2-2-11 1T, 3.1 V vdimm, 1.7 V vdd
- SuperPI 32M completed in 33 min 53 sec

 

[Adrayic]

CPU: San Diego 3700+ CABGE 05/15 VPMW
Memory: G. Skill 4400 LE 2 x 512
Motherboard: DFI LanParty UT NForce4 Ultra-D (Oskar Wu 414-2 Bios)
Cooling: XP-120 w/ Panaflo M1A
OS: Windows 2000 w/ SP4

CPU: 2900 Mhz = 290x10, 1.69V (VERIFICATION)
Memory: 290 MHz, 2.5-3-3-6 1T, 2.9 V
SuperPI 32M completed in 25min 1 sec

 

[El<(')>Maxi]

CPU: FX-55 0509 DPMW
Memory: Mushkin LII 2 x 512 MB BH-5
Motherboard: DFI LanParty UT NForce4 Ultra-D BIOS 4.14.3
Cooling: Baker SS loaded @ -48°
OS: Windows Server 2003

- CPU: 3.21 GHz = 268 MHz x 12, 1.87 V
- Memory: 268 MHz, 1.5-2-2-3 1T, 3.73 V
- SuperPI 32M completed in 22 min 53 Sec



*editited load temp, 10° warmer with FX vs. SD

 

[El<(')>Maxi]

Ran this for comparsion to my SD 4000. Same board, same Ram & settings exactly, just switched the CPU. Only difference is the PI version. It appears FX is faster and still king.....at least in this benchmark.

 

[Robert]

Only 22.53m? That is very very fast FX55.

 

[Robert]

This is 100% complete overclock result. Please update.

CPU: AMD 64 3000+ CBBID at 2.43GHz 1.55V
idle: 38C load:45C
MEM: KHX3200ULK2/1G TCCD C2.5-6-3-3-T1 2.65V
CPU:MEM 270:243
HTT 3x

 

[hitechjb1]

Ran this for comparsion to my SD 4000. Same board, same Ram & settings exactly, just switched the CPU. Only difference is the PI version. It appears FX is faster and still king.....at least in this benchmark.

Your two FX-55 have highest CPU clock and lowest SuperPI 32M run time, but in terms of the performance measure per CPU clock measured by

performance_per_CPU_clock = 1 / (runtime * CPU_frequency)

the two FX-55 are behind

San Diego 3700+ ......... 2900 MHz ..... 25.02 min ..... 13.78 ..... Adrayic
Venice 3000+ ............ 2700 MHz ..... 26.91 min ..... 13.76 ..... HousERaT
Winchester 3200+ ........ 2450 MHz ..... 29.88 min ..... 13.66 ..... SilverFire4
FX-55 ................... 2800 MHz ..... 26.24 min ..... 13.61 ..... Sucka
FX-55 ................... 3217 MHz ..... 22.18 min ..... 13.59 ..... El<(')>Maxi
FX-55 ................... 3180 MHz ..... 22.88 min ..... 13.59 ..... El<(')>Maxi
...
San Diego 4000+ ......... 3180 MHz ..... 23.48 min ..... 13.39 ..... El<(')>Maxi


All run are reported using the SuperPI 32M mod, except the SD 4000+ (non-mod version).

For details of these numbers and analysis, refer to
Different core performance results in SuperPI 32M
Analysis of different core performance in SuperPI 32M

Observation: One possible reason is that the memory setups from Adrayic and Sucka are based on TCCD running at higher frequency of 295/280 MHz respectively with 2.5-3-3-6 1T compared to your setup of 265/268 MHz tight timing 1.5-2-2-3 1T.

You may want to tweak the runs using TCCD based memory running around the 300 MHz level or the normal 2-2-2-5/6 1T setting for BH-5/UTT to see whether it helps the performance per clock number, and further lower the SuperPI 32M run time.

There were also OS variation among them, from W2K SP4, XP SP1/SP2, Windows Server 2003, have you tried W2k and Win XP.

 

[Gautam]

Your two FX-55 have highest CPU clock and lowest SuperPI 32M run time, but in terms of the performance measure per CPU clock measured by

performance_per_CPU_clock = 1 / (runtime * CPU_frequency)

the two FX-55 are behind

You didn't take into account his San Diego 4000+ run of 23 m 29 sec, vs his Clawhammer FX run of 23 m 7 sec at exactly the same settings. This is a significant difference, especially when you consider the only difference is the core revision.

To control one variable at a time, let's not consider which memory type is better just yet.

 

[deathstar13]

ok i swear im gonna make a path threw some winchester scores
i knew my mem timings were killing me so i bought some ocz value vx and scored some 2.0-2-2-5 goodness.but i did loose some mhz also in this as my mem controller isnt liking such high htt and low timings together.

i will get it faster as im burning the cpu more and the controller.the memory is just starting to strech its legs so im good to go!.

 

[rdrash]

new score today

CPU: ADA4000DAA5BN, CABHE, 0515SPMW (1235306D50216)
Memory: G.skill Pc4400 LE (2x512) 277Mhz, 2.5 tCAS, 3 tRCD, 3tRP, 7tRAS, 1T cmd rate, 2.9volts
Motherboard: DFI Lanparty UT NF4 Ultra-D, bios 4/14-3
Cooling: Stock HSF
OS: Windows Xp Professional x64 Edition Build 3790(service pack 1, v.1289)

CPU: 2770 MHz (277 x10)
Memory: 277 MHz, 2.5-3-3-7, 2.9v
SuperPI 32M completed in 26 minutes 10 seconds

Screen shots

 

[deathstar13]

raised my mhz and lowed timings and knocked it down too 32.00
this all i have left in my rig at the moment.better cooling may reap more results so mabey ill look into that.

hope i dont mess up your graphing posting new results.