KT266A vs KT333

[Maximus Nickus]

The sucess of the 8KHA+ might be down to PCI dividers or it might not be, Mushkin Rules! over at AMDMB.com had all his PCI cards (Modem, NIC, Sound, Vid) working at over 205FSB but the dividers might still not be there.

It could be down to many other things, for all we know maybe at different FSB's different Latencys, different Voltages or even different settings could be activated to attain that stability, maybe this is why Epox won't tell us why it can handle such high FSB's due to other people copying its designs.

Or it could be down to the board being so perfectly done (occasionally there will be a product so perfectly made it never goes wrong, lasts forever with incredible speed and stability, like a few old cars which just "feel" so well defined) that at high FSB's the co-ordination between the chipset the PCI's is spot on, We'll never know for sure but we can always wonder


M_N

 

[Malakai]

i love my 8kha+

if i get Samsung pc2700 and can max it at 200 on my 8KHA+, than i may get the 8k3a just for 10 more mhz on the fsb, unless epox will make it go higher in a bios update. we 8KHA+ users really need another one of those

-Malakai

 

[Maximus Nickus]

One of the 8KHA+ BIOS's goes to 210FSB< and WSPCredit (I think) was used to hit 235FSB on one

 

[Samzik]

can one make device

Can one make device wich could force the pci and agp to stay in the default 33 / 66 mhz. I think its posible, but we would need a good engineer to do that, and i think it only depends of the via chip, hows the pci agp is handeld. So maybe ... i just say maybe what epox is doen is slowing down pci and agp, showing them runing incredible speeds, but in real life they run much lower... Can i be wrong yes i can, but i can be werry right. So we never know what the board really does with the speeds when looking bios or eny mesuring programs, unles we be test it by hardwareware, because soft is always playen tricks as we all know.

So what u want to acheve here is near posible due everybody does it with soft, u can only hope on the ppz honesty but how reliable are ppz in net. Everyone wants to say i have biger then u do ......

Am i wrong am i right its ur desision its what i think if its not the subject u need to be in the post erase it or tel me and i erase heres sum what u have to think about if uv done it its ok chill and go on

best regards from Estonia Sam

 

[Sonny]

Samzik - You have a good idea for keeping the PCI & AGP speeds down. Remember how the Japanese moded the KT133 to be able to run at 133MHz+ I just hope for your puposes that it will not be to expensive. I actually like it when the AGP bus is out of spec. since it improves performance. The test on the Main page of Overclockers.com measure the PCI bus directly from the PCI slots so there is no chance of it going wrong unlike software tests that could be wrong.

Members - C'mon now guys & girls! How about showing all the proof that we need to find out which one is best? The guidelines are clear at the start of this thread. Let's not turn this into another "My KT266A has a 1/5 blah blah" & keep our goal insight.

 

[Malakai]

im just waiting on this Samsung pc2700 so i can post the benches fully, i still cant get 166fsb
-Malakai

 

[Silver]

8x166

 

[Silver]

9x150

 

[Silver]

10x133 mem

 

[Silver]

8x166

 

[Silver]

3d 10x133

 

[Sonny]

Silver - What happened to the CPU & MultiMedia scores & the 150MHz/300FSB 3DMark? Could you also show us the scores right before the divisor cuts in, maybe at 163MHz?

 

[Malakai]

i dont think he was able to run at 150, in a PM to me he mentioned that. he said his vid card doesnt like 150fsb

-Malakai

 

[Silver]

Malakai is right, the vid card will run most but not all tests before the gpu gives it up. As for the other tests, there are so many hours in a day. Will get them up though.

 

[Silver]

8x166 cpu

 

[Silver]

9x150 cpu

 

[Silver]

10x133 cpu

 

[Silver]

Do remember that 9x150 is a hair higher than the others and this should show up.

 

[Silver]

8x166 Multi

 

[Silver]

9x150 multi