Two P4 xeons, so at the end of the day you will reach performance somewhere around $20 office Dell computer with 2GHz Core2 :], Geforce 8800 would be perfect.

>The ZOTAC 1080 Ti clearly has better theoretical performance, but the TITAN XP has

this is just paper stats, cant even reach it in synthetic benchmarks, wouldnt worry about any of it. Afaik they calculate it by multiplying theoretical values.

rasz_pl wrote on 2023-09-24, 11:30:

Two P4 xeons, so at the end of the day you will reach performance somewhere around $20 office Dell computer with 2GHz Core2 :], Geforce 8800 would be perfect.

>The ZOTAC 1080 Ti clearly has better theoretical performance, but the TITAN XP has

this is just paper stats, cant even reach it in synthetic benchmarks, wouldnt worry about any of it. Afaik they calculate it by multiplying theoretical values.

This system's probably going to end up costing north of $2000, to give me $20 performance! (I enjoy playing with old tech more than new tech though!)

It's currently fitted with a GTX 690, and that gives better performance than an 8800 in the system!

Once I'm finished with this, I'll run a comprehensive set of benchmarks to determine it's exact performance!

Here's how it's CPU performance compares with that of some of my other system:

In that case Titan all the way, it was almost two times more expensive at launch 😀 and matches the theme much better.

rasz_pl wrote on 2023-09-24, 12:01:

In that case Titan all the way, it was almost two times more expensive at launch 😀 and matches the theme much better.

I'm thinking the same thing!

I already have a Maxwell version in one of my other machines which I bought when it was released!

However, one of these days I'm going to revet this machine to its original state (Red and Black), by refitting the Corsair Dominator GT RAM as well as the ASUS ARES 5970, as it will be far more collectable in this state! (Huge downgrade in performance though!)

The ASUS ARES is simply the most well made and beautiful graphics card I've ever laid eyes on! (Its heatsinks are made of solid copper and the card weighs somewhere in the region of 2kg!)

Irinikus wrote on 2023-09-24, 04:33:

So my question is the following: In the case where these GPU's are CPU bottlenecked to the extreme, will the difference in memory throughput (as well as the higher CUDA Core count on the TITAN) be an overriding factor when it comes to real world performance?

Remember that the goal here is to get maximum performance out of this dual NetBurst system, so please don't think as to how these cards would compare in the more modern type of system (i7) that was intended to run them.

Seeing that GPU's are SIMD-type systems, I'm personally of the opinion that the higher CUDA Core count would be a factor in this case. (Rather than the speed at which each individual core runs.)

Both bolded fragments are 100% software dependent.
If you run SuperPI, you won't see a difference between GPUs.
Alternatively, if you plan to run 3DMark 01SE, it is more likely that card boosting more often will get the top score (because running above 2D clock may be optional in most games under this setup).

Unless you plan on testing at 4k/8k there is no way you will be able to tell the difference between GTX 1080 Ti and Titan Xp on such setup.
Keep in mind, that driver overhead becomes a thing if your CPU is too low core/thread (or plainly too slow) for NV GPUs.
I like to link driver overhead being too high, with card time at 2D clocks during 3D tests (very NOT correct way to do it, but easiest to understand).

Side note : About GPUs being "SIMD" you mentioned, I would argue that if you have more than one GPC in GPU, it's more of a MIMD-type than SIMD.

agent_x007 wrote on 2023-09-24, 18:30:

Both bolded fragments are 100% software dependent. If you run SuperPI, you won't see a difference between GPUs. Alternatively, […]
Show full quote

Irinikus wrote on 2023-09-24, 04:33:

So my question is the following: In the case where these GPU's are CPU bottlenecked to the extreme, will the difference in memory throughput (as well as the higher CUDA Core count on the TITAN) be an overriding factor when it comes to real world performance?

Remember that the goal here is to get maximum performance out of this dual NetBurst system, so please don't think as to how these cards would compare in the more modern type of system (i7) that was intended to run them.

Seeing that GPU's are SIMD-type systems, I'm personally of the opinion that the higher CUDA Core count would be a factor in this case. (Rather than the speed at which each individual core runs.)

Both bolded fragments are 100% software dependent.
If you run SuperPI, you won't see a difference between GPUs.
Alternatively, if you plan to run 3DMark 01SE, it is more likely that card boosting more often will get the top score (because running above 2D clock may be optional in most games under this setup).

Unless you plan on testing at 4k/8k there is no way you will be able to tell the difference between GTX 1080 Ti and Titan Xp on such setup.
Keep in mind, that driver overhead becomes a thing if your CPU is too low core/thread (or plainly too slow) for NV GPUs.
I like to link driver overhead being too high, with card time at 2D clocks during 3D tests (very NOT correct way to do it, but easiest to understand).

Side note : About GPUs being "SIMD" you mentioned, I would argue that if you have more than one GPC in GPU, it's more of a MIMD-type than SIMD.

Thanks for the info! I appreciate it!

Thanks for pointing out my error regarding SIMD!

Seeing that there won't really be a difference in performance between the two cards, I'll opt for the TITAN, as it's a double slot card.

My SATA RAID card can only sit in the top (Orange) PCI-X slot, as it'a the only 133MHz 64-BIT slot on the board, making it a tight fit for a triple slot card such as the ZOTAC 1080Ti!