« on: September 27, 2017, 09:25:30 PM »
Interesting point of view, but I'm beginning to have doubts about our current direction.
The X5000 has worse video card performance than the X1000, and it has less PCI-E lanes than the X1000.
And the X1000 itself only has about half the PCI-E lanes of a PCI-E PowerMac G5 from late 2005 (18 vs 32).
Tabor has Gen1.1 PCI-E lanes like the SAM460, and like the SAM only four are connected to the video card.
And Tabor's e500v2 cored P1022 cpu is 32 bit, unlike the X1000, the X5000, and the G5 which are all 64 bit CPUs.
Then there are the fpu issues, the G5 and the PA6T in the X1000 both had very competent standard fpus with AltiVec instructions, the P50XX CPUs in the X5000 have an adequate fpu but they are missing the AltiVec instructions, and the P1022 in Tabor has a stripped down fpu with non-standard instructions and again no AltiVec.
So since 2005, we've moved from a quad core 2.5 GHz 64 bit system with AltiVec in a standard FPU on a board that supported 32 PCI-E lanes to a 1.2 GHz dual core 32 bit system with a non-standard fpu on a board with what about 5 PCI-E lanes (might be a few more, I haven't checked).
And this is the 'progress' that's going to carry us into the future.
Somewhere Jay Miner is laughing his ass off.:hammer:
In this benchmark, the X1000 ivs X5000:
https://keasigmadelta.com/blog/warp3d-novas-performance-boost-partially-hidden-by-lazy-cow/And it is a good read, the Sam460 - with the amount of pci lanes of the Tabor, is not that far behind btw.
« Last Edit: September 27, 2017, 09:28:44 PM by Bennymee »

Logged
Amiga 500, 1200, 4000, Amigaone, Morphos, CyberstormPPC, Blizzardppc, OS4.x