People are dying. Entire ecosystems are collapsing. We are in the beginning of a mass extinction. And all you can talk about is money and fairytales of eternal economic growth. How dare you! – Greta Thunberg
Just a thought, the x1k uses DMA to speed things up, would help if you had an older card to see if if its faster. Wondering if DMA is working on newer cards.
Look, only one leg, count em, one! X1000/PA6T@1800MHz/2Gb/Radeon 4850
I got curious about your older RadeonHD driver 2.10/2.11 ? And tested all i got archived(2.18, 2.21, 2.22)but result where only .1-.4 off between drivers.
What type of memory you got? what other sorcery you got goin on under the hood? ;)
I had pretty much ruled out the radeon driver. One thing I noticed was I had commented the FE1 graphics lib as "slow" & am using the lib from FE & its comented as "fast". For the life of me, I can't remember why. You could try v54.156 from FE. No idea if the x5k has working DMA or even if newish cards will even work with it.
Look, only one leg, count em, one! X1000/PA6T@1800MHz/2Gb/Radeon 4850
Well, i have "gfx PA6T DMA enabled" in my serial log.
It *is* turned on, it seems.
I'm going to try with the FE gfx lib next
X1000 with Radeon HD7750 ith OS4FE1
graphics.library 54.226 (13.09.2016)
GPmark improved, version: 2018 January 12 Surface dimensions: 640 * 480 Surface is located in RAM Surface must be locked: false Display mode: fullscreen Blitting Test: 180.3 Plasma: 150.9 Rotozoomer: 166.6 Rotozoomer Near: 167.0 Rotozoomer Far: 160.5 Radial Blur: 76.5 3D Bunny: 42.4
graphics.library 54.156 (19.10.2014) - from FE
GPmark improved, version: 2018 January 12 Surface dimensions: 640 * 480 Surface is located in RAM Surface must be locked: false Display mode: fullscreen Blitting Test: 197.5 Plasma: 166.3 Rotozoomer: 180.9 Rotozoomer Near: 176.9 Rotozoomer Far: 181.0 Radial Blur: 80.1 3D Bunny: 44.4
So the old gfx.lib *is* faster, but not to the extent it's showing for you.
Edited by Raziel on 2018/1/13 9:44:05
People are dying. Entire ecosystems are collapsing. We are in the beginning of a mass extinction. And all you can talk about is money and fairytales of eternal economic growth. How dare you! – Greta Thunberg
"So the old gfx.lib *is* faster, but not to the extent it's showing for you."
I still wonder if DMA is working right with newer cards, it was implimented before we had drivers for high end cards. Still waiting for someone with low #s to swap in a low end card to see what happens. I did pull my two 2GB mem sticks for two 1GB sticks, don't see that making any differece.
Look, only one leg, count em, one! X1000/PA6T@1800MHz/2Gb/Radeon 4850
Something is very wrong here...maybe it's the cards that differ?
People are dying. Entire ecosystems are collapsing. We are in the beginning of a mass extinction. And all you can talk about is money and fairytales of eternal economic growth. How dare you! – Greta Thunberg
Something is very wrong here...maybe it's the cards that differ?
Nothing is very wrong here, step away from the bench mark, move along ango about your normal business.
You need more convincing? Run a proper gfx bench card bench mark such as Hans GFXbench and see that your new cards are way faster than your old.
Don't run a poorly design CPU based benchmark rendering in CPU to **16**bit memory buffer that then needs to be converted to a bitmap and shoved across the PCI interface as fast as possible.
What you've discovered is that by some chance older cards are better optimised for converting 16bit raw gfx data in ram into bitmaps on screen. That could be because the end resultant data is differently endian requiring shuffling of the data or some other reason.
But when design actual applications which willl run on 32bit screens you actuall use 32bit data and get that data onto the card as soon as posible so that the GPU can work on it not the CPU.
@Sundown
DMA is certainly working on the newer cards, but curiosly switching it off in this test makes no difference on my card, most likely reason, the CPU based data conversion mentioned above is preventing DMA from being useful.