[bjames]

I cant seem to find benchmark data that compares the performance between a Dual R12K 400 mhz to a Dual R14K 600 mhz setup. Appart from the boost in Mhz, im more courious about the performance differences of the R12K vs R14K assuming the same clock frequency.

Just wondering if upgrading a dual R12 400 to a Dual R14K is worth money/effort. System is primarily used for May 6.5 using the V12 gfx card.

[Anick82]

I remember reading somewhere that the R1*K series were mainly R10K processors enginered to sustain higher speeds, with little optimization differences. So you can make a close prediction just comparing the total MHz. In your case 800 vs 1200.

Use this formula, while "X" is your actual rendering time, to obtain a rough estimate of the performance with the dual R14K ("Y").

X - (X * 0.333) = Y

60 minutes - (60 * 0.333) = 40 minutes

I followed the same steps in the past and the results got pretty close to the real benchmarks I found later.

[bjames]

I read somewhere suggesting the R14 is still based on the R10K architecture, how shrunken down to allow more "?" things to happen. What I was wondering if the advancements in the R14K are worth the upgrade.

The pricing on a dual r14K processor is still very high and my plan is to use this for a dedicated Maya 6.5 (and other 3D apps), however looking at the Alis Maya System qualification specs from Autodesk's website, the r14K Octane systems appear to have more issues reported compared to the R10/12K with both using the V12 gfx.

Part of the attraction for me to upgrade is having a maxed out system, but I dont want to do it if it introduces other problems.

[Anick82]

I read somewhere suggesting the R14 is still based on the R10K architecture, how shrunken down to allow more "?" things to happen. What I was wondering if the advancements in the R14K are worth the upgrade.

The pricing on a dual r14K processor is still very high and my plan is to use this for a dedicated Maya 6.5 (and other 3D apps), however looking at the Alis Maya System qualification specs from Autodesk's website, the r14K Octane systems appear to have more issues reported compared to the R10/12K with both using the V12 gfx.

Part of the attraction for me to upgrade is having a maxed out system, but I dont want to do it if it introduces other problems.

The raw performance would be like having a third R12K at 400MHz in your current setup, is not a big jump for the price and you already have the second best option in an Octane. The upgrade is not worth in my opinion, except if you find a bargain.

I didn't know about the R14K+V12 issues. It would be a good idea to list all the Octane compatibility problems in a thread. Yesterday I was fighting with two 2GB DIMMS from different brands that were not compatible in a 4GB configuration...

[bjames]

Thats the info I was kind of looking for. It sounds like the boost may be in some areas, but not significant. The Cost of a dual R14K is pretty high, so looks like I will stick with the R12K.

[Irinikus]

My Octane 2 with dual R14K’s@600MHz rendered test. blend in blender 2.45 in 3 minutes and 40 seconds, maybe someone with a dual R12K@400MHz can run the same test on the same version of Blender, to see how it compares.

[Irinikus]

Here are some results from Ian Mapleson's site showing blender test results between various SGI machines. He used blender 2.44 for these tests, which renders faster than 2.45, hence fourth, my machine's result is slightly slower than the result shown here.



According to these tests, the dual 600MHz machine is 1.5 times as fast and the dual 400MHz machine. (Therefore it is exactly a function of the CPU frequency in this case)

[Jammaster]

The dual R12K 400 vs quad PIII 500 is impressive !

[Irinikus]

The dual R12K 400 vs quad PIII 500 is impressive !

It is rather!

[bjames]

Looks like the overall render time performance increase correlates with the increase in processor speed MHz. So I would it appears the real gain is in the proc speed.