Memory Bandwidth Calculation and 2 other questions...

Hi,

I need a notebook for 3d modeling cars, bikes, boats etc... I'm looking at the T61p or M6300. The HPs are way more expensive...

From notebookcheck.net -

The T61p's FX 570m is 256 MB DDR3, clock rate Chip/VRAM: 475/702MHz.
The M6300's FX 1600m is 256 MB GDDR3, Core: 625 MHz, Memory: 800 MHz, 512mb with turbo cache

First - DDR3 and GDDR3 are the same right?

Second - Will the extra 256 mb turbo cache make any difference?

Third - How do I calculate the memory bandwidth of each card? because the nvidia website says 22.4GB/s for 570 and 25.6 GB/s for the 1600... but I've read some comparisons where the 570's was claimed to be only 12.8 GB/s (with DDR2) So, I'm kinda confused...

Thanks!

PS: Do you know what the shader clock means? and what they are for the above 2 notebook cards?

 

1st: The T61p is mislabelled, it's also using GDDR3 memory.

2nd: Doubt it, buswidth of both cards should be 128-bit so the extra 256 turbocache will only grant a small improvement in higher resolutions.

3rd: No idea, have you read Chaz's GPU guide and Meaker's chart for GPUs yet?

 

Thanks harley,
I just read through that guide... It has sorted out a few other questions I had, but didn't have any memory bandwidth calculation... Anyone have any idea?

 

Calculating theoretical bandwidth is simply multiplying video memory speed by video memory bus width and then multiplying by 2, because of the DDR (=Double Data Rate):
FX 570m: 702e6 * 16 * 2 = 22.464 GB/s
FX 1600m: 800e6 * 16 * 2 = 25.6 GB/s

Note that 128 bits = 128/8 bytes = 16 B
(1 byte = 8 bits)

That 12.8 GB/s is propably for the 14,1" version of the T61P where FX 570m has only 64 bit memory bus.