If this is your first visit, be sure to
check out the FAQ by clicking the
link above. You may have to register
before you can post: click the register link above to proceed. To start viewing messages,
select the forum that you want to visit from the selection below.
theoretically it could be possible (I think), but I don't think it's practical in use, and it probably would increase the cost of the card more than would benefit the increase in performance.
This would only allow higher throughput to the mainboard-RAM - and since todays mainboard's RAM can't even saturate AGP4x, this solution would have really no benefit, only bad effects: taking up one PCI-slot, misc incompatibilities (mechanical and on the software-side),...
BTW, since PCI has a lot lower throughput than AGP4x, you'd get nowhere near double throughput even if your system RAM could cope with the needed bandwidth.
The bandwidth-problem with current (esp. NVidias) gfx-chips is with the gfx-cards onboard-RAM. Here, having a wider path (i.e. a 256 or even 512bit wide path to the RAM) or the use of embedded RAM would surely help, but this is not feasible, at least not now at reasonable cost.
Or you could change the gfx-chips rendering-process to consume less bandwidth(tile-based rendering, HSR), which is probably the way to go.
[This message has been edited by Indiana (edited 20 February 2001).]
<font face="Verdana, Arial, Helvetica" size="2">Originally posted by Muted: I tought cards were limited because of the agp bandwith and all ??
i was reading something about the xbox beeing faster than a pc , because the Agp bus isn't fast enough
</font>
No. If you do some benchmarks of AGP2x vs. AGP4x you don't see any real performance-gain at all, even the difference between AGP1x and AGP4x is hardly worth the effort.
This is because the system RAM can't cope with AGP4x's bandwidth-need - this might be different in a P-IV system with RAMBUS, of course.
Comment