alabtrosMyster said:
For edram, esram, or any other type of buffer, even CPU and GPU L1 and L2 caches while we are at it, why not just ADD or multiply the bandwidth of all the busses into a single big number? because a 32MB buffer, even if it is almost as fast as the complete 8GB GDDR 5 the PS4 has does not make the xb1's 8GB of DDR3 as fast as the superior solution, at best in some very specific circumstances it may come close, on average probably not so much, but it will need to be flushed and refilled every few milisecond... this is not a problem if you only 1 bank of very fast memory. So in the end, in some cases if will make things a lot faster than if the gpu only had the DDR 3 RAM, in other cases it will barely make a difference, so you can't just add it up like that and compare them, they're not the same thing. So what's the number you would put for the xb1, xbox 360 and Wii U? because they all have that sort of setup. |
I understand that.
but then leaving it out makes sense? Totally not. Either make a comparison that does work and if you cant then dont do it because leaving it out just makes the whole chart completely useless.
Thats what I ment with its like "only 3 cores" its a different situation yes but ignoring more than 3 x1 cores would make the whole chart pointless.
If you dont want to add it to the DDR3 ram at least make a seperate "bar" and do some explanation or whatever. I mean the chart compares 2GB of DDR2 ram on WiiU with 8GB of DDR5 on PS4? But not 32mb or whatever of edram? Seems like double standards. Either both does not make sense or it does.










