# Thread: Xtreme whatever the heck you want to call this

1. Originally Posted by tcclaviger
You sir sound like a perfect candidate for a 3930k, 64GB of DDR 2133, and "By Core" OCing so that you can push say all 6 cores to like 4.8 (5.0 is not at all guaranteed stable on all 6), and then use per core to push it 5-5.1 when using only 2 cores. For thermal reasons and CPU safety it makes far more sense than shooting for 5.0x6core, especially when the app is primarily single threaded as you stated earlier. You may not even need to push it so high, the 3930K at 4.5 is dramatically stronger than the 9550 at 3.2.

Also, for this kind of work load I would stay far away from AMD.

Also, I am not sure you really want to put that swap file on the SSD, sounds like it will be doing an insane amount of operations on it, and at 20gigs, I would expect it drastically effect the lifespan of it. That is why I suggest the massive 64gb ram for a 2011 platform.

As an alternative, can you possibly use a different application that doesn't lean so heavily on the CPU for rendering and pushes more of it to the GPU(s)?

Perhaps something like this:
Attachment 126865
Can't really see the picture all that well, but I get what you're saying.

Definitely AMD is out of the picture.

I didn't really spend a great deal of time testing the 3930K loaner when I had it for CAD applications.

And re: SSD/swap/RAM

Well, unfortunately, Windows being Windows, you have to have a pagefile regardless of how much RAM you can. Yes, you CAN disable it, but I've tried that before and it really starts to behave weird/funky when you really start putting a load on it.

But otherwise yes, lots of RAM --- you can never have too much RAM. :oD

And I'm actually eyeing one of the MSI boards that supports 128 GB of RAM for the 3930K. Speed at that point isn't nearly as crucial as actually having it (since RAM will always be faster than even the fastest SSDs). And Kingston's one of the few manufacturers I know of that actually HAS 16 GB single modules (DDR3-1333) at $188 a pop. So, that'll be like close to$1500 in RAM alone. But the difference is, I will actually have something that can make use of all 128 GB of RAM (which is something that can't always be said for people that buy that much RAM).

And because it's mostly single processor applications (but still multi-threaded), I think that going with a 4-core 3770K would be plenty sufficient at 5 GHz.

The simulation nodes can be the ones that have 16-cores, and 64 GB of RAM.

And I'm using SSD as swap because unless I'm using a RAMdrive (which I've tried putting the swap onto that and it doesn't always work (depending on the RAMdrive/driver)), otherwise swap performance is actually one of my HUGE bottlenecks. Done a test before with a simulation where if I solve it using as little swap as possible (iterative solution) - I can get the result in an hour (and the system had a 7.2krpm SATA drive). If I do it with using maximum swap (direct solution), it takes 10 hours for the same problem, with the exact same solution.

I don't remember the test results that I did on the system that we had that had 128 GB of RAM (quad Socket G34 Opteron, 48 cores) min. swap vs. max. swap. (That one had a 240 GB OC Agility. It guy ordered the wrong SSD.)

Point being - swap = HUGE bottleneck for me.

It's the ONLY reason why I bought SSDs.

2. Keep in mind 3770k @5 is gona be A:hot as hell B:very hard to cool otherwise sure.

If swap is killing you, 4x WD Velociraptors in RAID 0 will be more reliable than an SSD, and very effective for it (also not terribly expensive)

3. Originally Posted by tcclaviger
Keep in mind 3770k @5 is gona be A:hot as hell B:very hard to cool otherwise sure.

If swap is killing you, 4x WD Velociraptors in RAID 0 will be more reliable than an SSD, and very effective for it (also not terribly expensive)
Yea, I read about that. But I've also started to read that people are saying not to worry about the high OC temps though. so *shrug* I dunno.

The last time I benched the SSDs, it was swapping at about 68 MB/s.

I think that due to the highly asychronous nature of swap, assuming perfect linear scalability, a single WD Raptor can only do about 5 MB/s (async). *4 - that'll still only be 20 MB/s tops. Minus RAID overhead. And it consumes more power, and generates a LOT more heat (compared to a single SSD).

The best that I can do with 2.5" 10krpm SAS drives is 8 MB/s (each) on swap, which means that even if I were to put four of them on RAID0, it'd still only be HALF the performance of an SSD.

4. Don't worry about the life span, even writing 2 TB per day you should get at least a year out of it. SSDs aren't THAT fragile, most of the things we do date back to the stutter issue on the old JMicron controller.

5. Originally Posted by [XC] Oj101
Don't worry about the life span, even writing 2 TB per day you should get at least a year out of it. SSDs aren't THAT fragile, most of the things we do date back to the stutter issue on the old JMicron controller.
Well, I'm hoping that my SSD will last more than a year. lol...

I think that in ONE of the test simulations that I ran, I forget whether it was read, write, or both; but it was pushing a few TB of traffic through on the drive, almost all of it transient data.

Just out of curiosity, what happens after you burn through all of the P/E cycles on a drive anyways? Do you start losing capacity? Performance? Both?

6. Originally Posted by Liam_G
Just curious, why are you running simulations of a combustion engine for a hobby, why not build the engines? Especially when it will set you back 20k to do it.....

I only run simulations to get an idea of what will happen before i validate that testing in real life. I will eventually justify spending money for my solidworks hobby because in the end i need to upgrade my system, play games and i do build the stuff i design....

So are you building engines? :P because that would be cool, i used to draw plans for 2 stroke motorcycle engines as a kid and always wanted to build an engine from scratch, however small and simple it was, i even designed a model in Autodesk inventor when i first learnt CAD in highschool
CR4P! I'm late to the party. Again.

http://www.technologyreview.com/ener...nld=2012-05-17

File the homogenous charge (gasoline) compression ignition as "been there, done that (already)."

7. Popped the hood.

8. I was reading on Tom's Hardware that between the 3770K and the 2700K, that the performance advantages isn't nearly as great as one might think/suspect.

http://www.tomshardware.com/reviews/...k,3181-24.html

But since in this case, it's mostly a single-processor app, it IS multi-threaded, but not multi-processor capable (no slave processes); would it be better for me to to OC the 2700K to 5 GHz or should I start with the 3830 and push THAT to 5 GHz?

Thoughts?

9. 3820
2700
3770

If you get similar quality chips and push them all to reliably high OCs, 4.8 for SB and SB-E, and 4.6, you will have almost identical performance. In certain tasks each one out performs the others. Cost/Heat should be more the deciding factor than absolute performance.

3930k
3960x

I think the X is a waste of money for anything other than competitive benchmarking. The K has been spotted performing better clock for clock in a handful of tests, however, from anecdotal evidence seems to consistently clock higher (could be due to sample size since there are probably far more K chips out in the wild than X chips). Either one will absolutely smoke the first 3 in pretty much every single task clock for clock and if they dont pass it, it will be within a percent or two.

5 Chips, all great pick your poison