maverik-sg1
New member
Okay guys just running at 206x16 (3.293mhz) on the FX60, reduced clock to try and give some comparisons agains the Yonah chip:
Here's my result:
Here's my result:
FX60 Versus Yonah Superpi32m
- Thread starter maverik-sg1
- Start date
maverik-sg1
New member
This is Stummermeister results from XS:
Rumours of cache inhinitors and alike - may be more appropriate for Conroe - certainly not the case with this beauty - although I think we are seeing that this is currently the max overclock that anyone can achieve so far (correct me if I have assumed wrong).
Even so clock for clock - tis better than FX60 :O
Rumours of cache inhinitors and alike - may be more appropriate for Conroe - certainly not the case with this beauty - although I think we are seeing that this is currently the max overclock that anyone can achieve so far (correct me if I have assumed wrong).
Even so clock for clock - tis better than FX60 :O
Highland3r
New member
Don't mean to sound funny/offish or whatever but this isnt exactly a like for like test...
Memory speeds are way off (assuming CPU-z's reading right which it appears to be for a 133 divider) FSB's are different, CPU speeds are different....
If you run the ram sync on your FX, it'd be interesting to see how evenly matched they are (again thats assuming CPU-z's reading the divider right...)
The Yonah's seem to suffer with the shared cache too, comparing single vs dual Pi runs, single is faster on Yonah. Carrying out the same task using A64's (or other chips with dedicated cache per core) shows a MUCH MUCH smaller performance difference (if any at all)
As http://www.xtremesystems.org/forums/showpost.php?p=1421461&postcount=127 shows
Thats probably the only weakness in the architecture, with dedicated cache per core it'd be faster than it is, but for mobile use the shared cache makes MUCH more sense
Memory speeds are way off (assuming CPU-z's reading right which it appears to be for a 133 divider) FSB's are different, CPU speeds are different....
If you run the ram sync on your FX, it'd be interesting to see how evenly matched they are (again thats assuming CPU-z's reading the divider right...)
The Yonah's seem to suffer with the shared cache too, comparing single vs dual Pi runs, single is faster on Yonah. Carrying out the same task using A64's (or other chips with dedicated cache per core) shows a MUCH MUCH smaller performance difference (if any at all)
As http://www.xtremesystems.org/forums/showpost.php?p=1421461&postcount=127 shows
Thats probably the only weakness in the architecture, with dedicated cache per core it'd be faster than it is, but for mobile use the shared cache makes MUCH more sense
maverik-sg1
New member
No memory timings are actually higher than that - DFI allows you to 'increase' the clock speed of the RAM using a multiplier, CPUZ does not understand this and thus reports it as a divider.
The RAM was running around 256Mhz versus a 206Mhz HTT speed of the FX60.
Mav
The RAM was running around 256Mhz versus a 206Mhz HTT speed of the FX60.
Mav
