All posts relating to Maxwell Render 1.x
By sandykoufax
#176668
Ahhhh, woodcrest 2core's result (1h24m56s) is a bit disappointed. :(

I know that Xeon 5160 is top model of woodcrest.

It just faster only 16m than Xeon 3.8GHz' 1h40m. :cry:
By adamwade
#176676
Hey Tom,

I can understand that time and SL are not to be compared between V1 and V1.1, but I thought Benchmark should be a fair comparison? Is this wrong as well?
User avatar
By Thomas An.
#176682
adamwade wrote:Hey Tom,

I can understand that time and SL are not to be compared between V1 and V1.1, but I thought Benchmark should be a fair comparison? Is this wrong as well?
Time should be the only criterion. In other words you render the same scene between the two engines for the exact same amount of time.

However, as Tom mentioned, benchmark numbers and SL numbers do not correspond between different engine versions and should not be used for comparison.
By adamwade
#176683
Yeah, but that means we just compare the 2 images and see which one looks best? There has to be a an SL or Benchmark to compare. OR, I guess we could all test until SL 25 limit then see who gets there first ..... some of us may not be alive by then to post our results :wink:
User avatar
By Thomas An.
#176686
adamwade wrote:Yeah, but that means we just compare the 2 images and see which one looks best? There has to be a an SL or Benchmark to compare. OR, I guess we could all test until SL 25 limit then see who gets there first ..... some of us may not be alive by then to post our results :wink:
I know ... but still the numbers do not correspond, so going by SL or benchmanrk numbers is no better than going by eye.

The only time we can use SL or benchmark is when doing CPU tests and render the same image accross several PCs

Doing a cross-engine is tricky and can be tedious. There was an attempt here, but it is involved and it needed to sample "regions" to avoid the noise patterning effects.
http://www.maxwellrender.com/forum/view ... t=isolates
By adamwade
#176705
Here is the result from our newest machine at work:

Dell Precission 490
Intel Xeon Core Duo 3.0 GHz 5050 (X2) = 4 cores
2GB RAM
XP Pro

Maxwell V1.0
Benchmark= 38.76
Time = 108 min.

Console read: Processors: 8 logical/ 4 physical

I am a little disappointed that it didn't even match the Mac G5 Quad core, but then I am also smiling because I love Macs 8)
User avatar
By beppeg
#176744
:shock: very strange :shock:
like the JDHill's Dual Xeon 3.8 GHz HT @ 4.0 GHz :?:
But andrebaros, in another post, gets 84' instead of 108' :!:
By sandykoufax
#176789
begpeg, andrebaros's new xeon is not the same with JD's.

He said that his machine (Xeon 5160) is woodcrest, not irwindale or dempsey.

IMO, andrebaros' benchmark result of 1st page list is not woodcrest, maybe his old machine.

Anyway, result of woodcrest is disappointed. :?
Last edited by sandykoufax on Wed Aug 09, 2006 3:48 pm, edited 1 time in total.
User avatar
By beppeg
#176810
:oops: sorry sandy for the confusion :(
I was talking about the two Intel Xeon Core Duo 3.0 of andrebaros and adamwade, now I see that they have differnt numbers, 5050 and 5160 and also not the same number of cores :oops:
User avatar
By beppeg
#176811
Another thought :!:
adamwade, have you set the numer of th = 4 in the mxcl?
User avatar
By glassbathroom
#176825
sandykoufax wrote:Ahhhh, woodcrest 2core's result (1h24m56s) is a bit disappointed. :(

I know that Xeon 5160 is top model of woodcrest.

It just faster only 16m than Xeon 3.8GHz' 1h40m. :cry:
I'll run it on my Quad Core 2.66 Xeon (5150) Mac Pro, when it gets here.
By sandykoufax
#176839
Yes, adamwade's benchmark is stange.

quad core xeon's mark is slower than athlon x2 dualcore. :?:

Does anybody has Core 2 Duo (E6600 or others)?

I heard that Intel's conroe and woodcrest are almost same core except FSB.

Then, E6600(2.4Ghz) would be slower than woodcrest 5160 (3Ghz).

If so, it's not attractive than athlon x2. :?

We needs more benchmarks. :!:
By adamwade
#176961
beppeg,

I will check again that my threads were set to 4, but in my post I said that the console read:

Processors: 8 logical/ 4 physical

I thought this meant everything is working.
By adamwade
#177062
:oops: Hold everything !

Yes my render threads were set to 4, but that's wrong! I was testing a 4 core machine so I realized that it should be set to 8 threads.

WOW! what a difference: so here is the real results......

Dell Precision 490
Dual processor (4 cores)
Xeon 3.0 GHz 5050 dual core
2GB RAM
Windows XP Pro

Test Scene:
Benchmark=62.14
Time= 69 min's
User avatar
By glassbathroom
#177084
That's a bit more like it! Well done.
  • 1
  • 6
  • 7
  • 8
  • 9
  • 10
  • 12

> .\maxwell.exe -benchwell -nowait -priority:[…]

render engines and Maxwell

You could be right about AI, but actually I prefe[…]