AGP is 66mhz... Have you, yourself read the spec?
AGP4x is 66mhz 32bit quadpump.
And are you calling people reporting 130-200meg read under OpenGL lyars?
Example:
nVidia Quadro DCC / P4 Xeon 1.5 GHz x 2
OpenGL Write: 482.03 MB/s Read: 157.60 MB/s
Email: d a n i e l @ e y e o n l i n e . c o m.
if you want to prove yourself wrong.
Stephan
The Real3d / Intel 'partership' resulted in the starfighter card. This card was VIVO ready and supported DMA both ways across the AGP bus.
You could download from the card at 20meg second effortlessly using a pentium classic.
Yet a 2.5ghz pentium4 cant get much more then 10meg second with a geforce4 under XP using 100%
of the processor resource.
Interestingly, OpenGL on similare HW can deliver 20x that performance.
No new HW specification is needed... updated drivers are needed.
Stephan
Its no HW issue. If it was it wouldn't explain how people get 200meg second readback using OpenGL
or how some are getting 80-130 meg readback under win9x. (This apply only to nvidia drivers, I'm not aware of any other driver that perform readback using DMA)
Stephan
This does not explain why the benchmark work 8 to 13time faster under Win9x and people posting result of 130-200 meg scond using OpenGL.
Its a software issue. No need to invent elaborate HW hacks.
Stephan
Can anyone confirm that the AGP spec is not symmetric?
The fact remain... Nvidia card under OpenGL deliver 200meg second , yet most card (include nivida) deliver 15meg with DirectX and XP.
Note: We did see nvidia driver deliver 80-130meg second under W98.
Their is a big gap betwen 10meg to 200meg.
Also another factor is that in that benchmark 100% of the processor is used (driver use CPU loop VS DMA) And its very probable that OpenGL using DMA and freeup the processor. So 200meg second 'only' but with little procesor usage.
Double bonus!
Stephan
Its a Dx8 benchmark and use the API/Driver
to copy the buffer.
The benchmark was actually seen delivering 80-130meg second on the win9x platform.
Its a driver issue for sure, but might also be a W2k/XP (in combination with the directx) API limitation.
This a good reason for people to stick to using OpenGL. But would be nice if DirectX can deliver the same performance nontheless.
Stephan
Its not the benchmark test code that copy the data with the CPU, its all in the driver/API hands. Its all done in one API call.
This same API call for example easely deliver 80+meg second on nvidia driver and W9x.
Its all about the driver implentation....
Stephan
Nvidia even implement DMA transfer in their OpenGL driver, and their W9x directx drivers.
GPU are allot more flexible then you can imagine. Being close minded doesn't help the industry move forward and be inovative.
Millions of user would prefere to do video work in realtime/faster then realtime. Thats just one market/application for putting DMA support in drivers.
The 4+ million card sold every month dont end up in gammer box. Other market do exist...
This is wrong. GPU are designed to do this also.
Each core is VIVO enabled, from the TNT to the Geforce4.
When you capture video on a Geforce2 or a AIW radeon the video is capture passing thru the AGP buss using DMA.
Their little load on the CPU/system resource.
Those video surface after all can be compressed using software mpeg2 in real time.
Whats missing is the driver path that also perform this for graphic surface, not just video surface.
I seing a bunch or report from people seing 100meg under 9x but 10meg under Xp/W2k with nvidia HW...
Its all about drivers.
Stephan
Slashdot Mirror
AGP is 66mhz... Have you, yourself read the spec? AGP4x is 66mhz 32bit quadpump. And are you calling people reporting 130-200meg read under OpenGL lyars? Example: nVidia Quadro DCC / P4 Xeon 1.5 GHz x 2 OpenGL Write: 482.03 MB/s Read: 157.60 MB/s Email: d a n i e l @ e y e o n l i n e . c o m. if you want to prove yourself wrong. Stephan
The Real3d / Intel 'partership' resulted in the starfighter card. This card was VIVO ready and supported DMA both ways across the AGP bus. You could download from the card at 20meg second effortlessly using a pentium classic. Yet a 2.5ghz pentium4 cant get much more then 10meg second with a geforce4 under XP using 100% of the processor resource. Interestingly, OpenGL on similare HW can deliver 20x that performance. No new HW specification is needed... updated drivers are needed. Stephan
Its no HW issue. If it was it wouldn't explain how people get 200meg second readback using OpenGL or how some are getting 80-130 meg readback under win9x. (This apply only to nvidia drivers, I'm not aware of any other driver that perform readback using DMA) Stephan
This does not explain why the benchmark work 8 to 13time faster under Win9x and people posting result of 130-200 meg scond using OpenGL. Its a software issue. No need to invent elaborate HW hacks. Stephan
Can anyone confirm that the AGP spec is not symmetric? The fact remain... Nvidia card under OpenGL deliver 200meg second , yet most card (include nivida) deliver 15meg with DirectX and XP. Note: We did see nvidia driver deliver 80-130meg second under W98. Their is a big gap betwen 10meg to 200meg. Also another factor is that in that benchmark 100% of the processor is used (driver use CPU loop VS DMA) And its very probable that OpenGL using DMA and freeup the processor. So 200meg second 'only' but with little procesor usage. Double bonus! Stephan
Its a Dx8 benchmark and use the API/Driver to copy the buffer. The benchmark was actually seen delivering 80-130meg second on the win9x platform. Its a driver issue for sure, but might also be a W2k/XP (in combination with the directx) API limitation. This a good reason for people to stick to using OpenGL. But would be nice if DirectX can deliver the same performance nontheless. Stephan
Its not the benchmark test code that copy the data with the CPU, its all in the driver/API hands. Its all done in one API call. This same API call for example easely deliver 80+meg second on nvidia driver and W9x. Its all about the driver implentation.... Stephan
They all support DMA capture over the AGP bus.
Nvidia even implement DMA transfer in their OpenGL driver, and their W9x directx drivers.
GPU are allot more flexible then you can imagine.
Being close minded doesn't help the industry move forward and be inovative.
Millions of user would prefere to do video work in realtime/faster then realtime.
Thats just one market/application for putting DMA support in drivers.
The 4+ million card sold every month dont end up in gammer box. Other market do exist...
Stephan
This is wrong. GPU are designed to do this also. Each core is VIVO enabled, from the TNT to the Geforce4. When you capture video on a Geforce2 or a AIW radeon the video is capture passing thru the AGP buss using DMA. Their little load on the CPU/system resource. Those video surface after all can be compressed using software mpeg2 in real time. Whats missing is the driver path that also perform this for graphic surface, not just video surface. I seing a bunch or report from people seing 100meg under 9x but 10meg under Xp/W2k with nvidia HW... Its all about drivers. Stephan