Maniaplanet Forum

yo, I said I have only canyon, I can not use this maps

canyon map with 6182 coppers on HQ take 7m23s.
Q9550s@3,4Ghz@avg: 40% and GTX670@1016mhz(no turbo)avg:21%

same map, but without building around with 2530 coppers on HQ take 1m19s.
Q9550s@3,4Ghz@avg: 45% and GTX670@1175mhz(turbo)avg:71%

cpu 40% means 1,6 cores of 4 and 45% means 1,8 cores of 4

I can test on these maps. I have Canyon.

When you are saying HQ, that means High and not Ultra, isn't it ?

Can you give me the Tm exchange ID (or direct link) for these map ?

yes, high quality.
I also used ultra, but after reinstall windows i didn´t set the checkbox for ultra
I have no clue whats special, I didn´t seen as much difference.

if u want my map, i uploaded it on my ml
I don´t use mx.

lolzone?lolmaps @ 10 planets
it´s the newest one: TM²Stadium LoL #74

btw. I didn´t saved the less blocks edition, only deleted for testing.

4 min 43 sec for the TM²Stadium LoL #74
With Core i7@3.4Ghz 8GB and ATI Radeon 7950

Ps : I firstly didn't understood that ML == Manialink xD

You Manialink is nice.

But the editor said me he didn't found 2 dds and 1 jpeg. No matter, I was able to compute !

We will compare on exactly the same map with my friend.

I have test with my friend on the same map.

I have posted the result here : http://forum.maniaplanet.com/viewtopic. ... 79#p140479

niarfman wrote:4 min 43 sec for the TM²Stadium LoL #74
With Core i7@3.4Ghz 8GB and ATI Radeon 7950

for me ati looks faster xD

niarfman wrote:Ps : I firstly didn't understood that ML == Manialink xD
You Manialink is nice.

ty

niarfman wrote:But the editor said me he didn't found 2 dds and 1 jpeg. No matter, I was able to compute !

ohh, thets bad, told he the file names?
could be helpfull to know

thx

Ok, the files my editor wasn't able to download are :

• Skins\Any\Advertisement\lolzone_adv2_1.dds (http 404)
• Skins\Any\Advertisement\lolzone_adv1.jpg (http 404)
• Skins\Any\Advertisement\lolzone_adv3.dds (http 404)

About the tests I made with my friend, Nvidia seems faster (45% in our case) with similar CPU Core i7.

I have also launched the computing with an other computer running a Core2 Q6600@2.4GHz - 4GB DDR2 - Radeon 5450. I need 30 minutes !!!

Do ok the clok is 2.4GHz and not 3.4GHz. But it's not 1GHz more that can make a such difference.

If my computer was faster it's mainly because i7 is very faster than a Core Quad. It's very good to reach 5 minutes with a Core Quad !

The difference with that PC is the GPU power and not the CPU power. And you can't seriously try to compare a Radeon 5450 with a Radeon 7550?
I mean, they are completely different systems, different publication years and price ranges.

The Radeon HD 5450 is on the VLIW5 architecture (40nm) while the 7950 is a high end performance card on the new GCN architecture (28nm). It's like comparing a VW beetle with a Mercedes SLK. Furthermore, AMDs rebranding process makes it even worse. Soon the 5450 will be available as a Radeon HD 8350. Therefore you need to compare everything with the same system. Else you can't say anything.

Indeed, but it was to show the benefit of GPU acceleration here.

The lightmap calculation seems to be optimized for Nvidia GPU and not for Radeon.

As I said, we have made a benchmark between 2 Core i7 and Radeon 7950 in a case and GeForce GTX 680 (faster than Radeon 7950) in the othe case. Configuration with GTX render within 11 minutes where Radeon needs 16 minutes. It sounds like the Radeon GPU is not used. So, in this case, using a Radeon 7950 or 5450 should be the same.

The best test would be to use The same plateform with 2 graphic cards ATI / Nvidia in the same range of performance and using the SAME MAP

I can also test an old GeForce 8400GS (Cuda compatible) and see if it's faster than 7950. So I will know if it's nvidia optimized or not

The game is more optimized for Nvidia, the studio works with Nvidia cards essentially.