by Mark Rubrico

 

Mini review of my li’l duron 700 that struggled with me.

 

CPU Tidbits

·         Duron 700 - 1,800 pesos

·         D700AUT1B

·         AKCA0109XPCW

·         Z87748605M

·         Locked @ 7x Multiplier

·         Cooled by Smallest HSF Combination available

 

 

 

SYSTEM SPECS THAT ARE RELEVANT :

·         Savage 3D 8MB ( around 115 Mhz Clock, Overclocked )

·         1 Realtek 8029 10Mbps

·         1 CNET 2100B 10/100Mbps

·         IWILL KK266 Revision 1.2

·         1x128 Apacer PC133 Ram – CAS3@133 L

·         1x256 Generic PC133 Ram – CAS3@133

·         3x80MM “power supply” 12V Fans

·         1 Duron 700 ( Default Voltage 1.6V ) – locked

·         Crappy Heatsink with 6100+RPM Fan Installed on Top

·         Radio Shack Thermal Paste

·         1x15 B Maxtor Hard Drive ( Supposed to be sensitive to high FSB, NOT! )

·         1x4x4x24 CR4804 ( Mitsumi, one of the best CDRWs evermade J )

·         1x50x ASUS CDROM drive ( the 40x is much better )

·         1 HANABISHI Desktop Fan 220V

 

 

 

HIGHEST FSB Attained : 2x155 ( 310 effective, DDR ) completely stable

·         did not bother to go higher, scared of frying my lan cards

·         PCI = 38.75 Mhz

·         AGP = 77.5 Mhz ( AGP4X enabled )

 

HIGHEST CPU Clock Attained : ~1100 Mhz

         

·         Highest Stable Clock attained, 975 Mhz, 6.5x150

·         Rock Stable @ 975, running Prime 95 right now.

·         Core Voltage used, 1.85 ( Sandra reports ~1.99V )

·         VIO Voltage Used, 3.3V ( Default ), Sandra reports ~3.4

·         Temperature @ Full Load ( Prime 95 ), with Desk Top Fan, 52degrees ( case 28 degrees )

·         Temperature @ Full Load ( Prime 95 ), without Desk Top Fan, 55 degrees ( case 32 degrees )

·         Temperature @ Full Load ( Prime 95 ), without Desk Top Fan, 58 degrees, cover on ( case 32 degrees )

 

HIGHEST Temperature Attained :

·         Core - 61 Degrees Celcius, did prime 95 without Fans ( ex CPU HSF )and went to CERTII, damn CERTII.. hehehehe..

 

 

MODIFICATIONS

·         “soldered” L1 Bridges of Duron ( unlock multiplier )

·         max core speed attained, 950 Mhz, 9.5x100 ( POST, flaky with WinME )

·         max Stable core speed attained, 910 Mhz, 140x6.5

·         max FSB 155Mhz

·         “soldered” L7 Bridges of Duron ( to raise Default Voltage from 1.6 --> 1.85 )

·         max core speed attained, 1100, 10x110 ( POST, flaky with WinME )

·         max stable core speed attained, 975 Mhz,  6.5x150

·         Left VIO Voltage @ Default ~3.3

·         Test For Stability include Sandra Burn In utility, Prime 95

 

 

Benchmark DETAILS

 

 

Notes :

         

·         The athlon would be faster due to the larger cache and perhaps because of its prefetch capabilities, if this happened to be a palomino/mp. But then again, this Duron costs only 1,800 pesos!!!!! hehehehe

·         Synthetic benchmarks suck… hehehe.

·         Scaling linearly, 25mhz ( this is a 975 Duron ) more in clock would make the Duron perform like a PIII ( Take Note, SiSoft Sandra is Heavily Optimized with SSE2, SSE optimizations, so that leaves real world performance to the Duron )

         

Hmmm… Processor duron 975 Mhz

Voltage 1.5-1.7, hehehe, mines@1.98, look at the last line!! Next!

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hmmm.. Iwill kk266, yup I flashed bios. That’s supposed to be a KT133A chipset. And maximum installable memory shouldbe 1.5G, Ram is CAS3@150

And yes, 4-way Bank Interleave is on, coz of the new VIA 4 in drivers. Apacer Ram worked at 150Mhz!!!! Generic Ram Worked at 150 mhz!!!!! Talo pa Micron KO! Hehehe..

 

Front Side Bus, 150 mhz. DDR! 300MHZ!!! Hehehe,

 

 

 

 

 

 

 

 

 

 

 

 

Notes :

·         At 150Mhz@CAS3 my rig is threatening the DDR ( PC2100, that’s bandwidth offered )  platform in terms of memory performance! If I had a stick that would do CAS2 at 150, then there would be absolutely no competetion from this particular DDR platform!  This is evident by the reference KT133As ( third row ) performance.. however the difference of CAS2 and CAS3 at any given clockspeed in real world applications, say… Quake 3, would be 2-3% at best.

·         The KT133 chipset should be avoided as its performance is really not on par with the KT133A. And the fact that it “only runs” at 200mhz FSB is really hurting it. ( KT133A = 266mhz, 33% anyone? ). 1.6GB of Bandwith vs. 2.1GB of Bandwith ( or 2.4GB if u can run @ 300Mhz Fsb like me! Hehehe )

·         HAHAHAHA, RAMBUS! That thing runs at 800 Mhz! Hehehe, they come with heatsinks too!

 

 

 

 

 

Additional Benchmarks.

-          there it is, the description of this so called cpu multi media benchmark.

-          This very “synthetic” since in actual software, the enhancements done aren’t as thorough as it is here… aka 3dmark 2001. the p4 will shine here. I think.

 

 

 

 

 

Conclusion/Delusions

 

The elusive 1Ghz crown, was well… elusive! Hehehe….

 

1800 pesos for 1,000,0000,000 Hhz ( 1 Ghz )? Not quite… but damn close! At 975 ( 6,5x150 ) this will outperform a 1Ghz ( 10x100 ) and even a 1Ghz ( 7.5x133 ) !!!! A <10 percent increase in FSB would not only negate the 25mhz difference in clock, but it will allow the 975 (6.5x150 ) to surpass the 10x100 ( and yes, even the 7.5x133 )in virtually all real world applications… well, most!

 

Warning if u ever do this, u must remember that u will void ur Duron’s warranty. I had to remove the warranty Sticker inorder to get to the L7 bridges.

 

Mark Rubrico

mryoso@yahoo.com

feel free to comment!