hardware speed... rendernode

jano · Post by **jano** » Fri Mar 21, 2008 8:57 pm

Hi all.
On your opininion wich is the performance of this machine only running indigo? I'll use this as internal server + indigo:

· Genesis rendernode GEREXE5345
· 01 rack 1u con alim da 650 Watt EPSA ad altissimo amperaggio specifico per Xeon e 7 ventole di raffreddamento
· 01 Mb Dual Xeon LGA771, (8) 8 FBDIMM, Audio, (2) GbE, SATA RAID, video certificata
· 02 CPU XEON 5345 Quad core 2.33 GHZ FSB1333 ACTIVE 4MB CACHE
· 04 Moduli memoria DDR II 667 ECC FB DIMM 1GB certificati kingston
· 04 HD WD Caviar SATA II RE 320 GB 16 MB cache classe RE garantito 1.2 milione di ore in raid
· 01 let dvd dual layer slim
· 01 win xp pro 64 bit

bye,
jano

Zom-B · Post by **Zom-B** » Fri Mar 21, 2008 9:06 pm

Performance will be very similar to a Q6600, check out this thread for some benchmarks etc.

Indigos Performance is based mostly on pure Mhz power, so performance of a PIII or a Xeon Quad core should be about the same for each Mhz (regarding the Xenon on 1 rendering thread).

dougal2 · Post by **dougal2** » Sat Mar 22, 2008 12:45 am

that is a dual quadcore xeon system. Expect performance to be roughly twice that of my q6600 - although your memory speed isn't as fast as mine, and that is another important factor (and is why my quadcore system runs 8 times faster than an older 3ghz p4, not 4 times faster).

Zom-B · Post by **Zom-B** » Sat Mar 22, 2008 1:10 am

dougal2 wrote:... although your memory speed isn't as fast as mine, and that is another important factor (and is why my quadcore system runs 8 times faster than an older 3ghz p4, not 4 times faster).

Hmm... I have old Rig with an AMD 2100+ (1733Mhz) on 512MB SD RAM, and a AMD X2 3800+ (2x2000Mhz) with 4GB DDR1 RAM, both systems scale pretty the same way for each Mhz...

Are you sure about the +100% boost only because of other (faster) RAM O.o

BTW: yes, now I see it too, 2 Xenons...

dougal2 · Post by **dougal2** » Sat Mar 22, 2008 1:18 am

well if i'm honest, i'm not entirely sure, but i'm not sure how to account for it.

It illustrates that there is _not_ a linear scaling with MHz.

4 x 2.4ghz = 9.6ghz,

9.6ghz = 3.2 x 3ghz

When in reality, the figure is somewhere between 7 and 8. It's probably due to faster FSB speeds, different bus architectures etc etc, but the one thing I do know is that the RAM is also twice as fast.

zsouthboy · Post by **zsouthboy** » Sat Mar 22, 2008 2:00 am

Also, comparing against a P4 based processor - the P4 had incredibly long pipelines, and was therefore _much_ slower on branchy stuff like Indigo than its MHZ indicates. (But it was great for sales, wasn't it, Intel?)

The Core 2 series has its roots in the P3, which doesn't have this problem.