• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Tips for new PC

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Certainly a good lineup; not sure about using the SSD, there are both benefits and cons, I'll explain later.

The processor definitely has the most bang for the buck; I certainly wouldn't recommend getting anything less than that, especially at it's price. See these charts for comparisons (Highest gain is +450 pounds for 16&#37😉; as you can tell, the i7-930 is quite a deal for its power.

Rendering time of (x) file in Adobe Premier Pro CS4
Rendering time of (x) file in Autodesk 3DS Max 2009

Excellent choice on Mobo. Asus has dropped in reliability over the last 5 years or so, but they're still fairly good and most of the issues are DoA boards. Just be aware you may need to pay for shipping 2 or three times on this board. Or maybe once; luck may shine on you.

For the parts you've specified the wattage comes to ~446 watts which, assuming standard 10% depreciation, your PSU will be able to support for at least 4 years, more likely 5-6 or more.

I'm not certain about the benefits of the SSD setup. Yes, it's very fast, but also pathetically small. Small enough that in less than 10 full length HD Video projects you'll fill the fscker if it's your primary drive. The solution would be to run an automatic backup to a RAID 1 using the F3's, which would be terribly slow since you loose like 20% of the drive's speed in RAID 1. If you ran the backup at night it'd be a moot point, but if you spend a lot of time pulling old data, you loose the benefit of the SSD.

I have about 140 pounds for your SDD, which worth almost 3 F3's. Let's compare a RAID 0+1 with 4 F3's against the SSD w/ a RAID 1 (F3's) as an archive.

Setup 1: Primary Drive: Redundant? Yes
Size: 2 TB
Avg Read: 220 MB/s
Avg Write: 180 MB/s
Archiving required? No

Cost: ~200 Pounds

Setup 2: Primary Drive: Redundant? No
Size: 60 GB
Avg Read: 285 MB/s (+30% spd)
Avg Write: 275 MB/s (+35% spd)
Archiving Required? Yes

Secondary Drive: Redundant? Yes
Size: 1 TB
Avg Read: 190 MB/s (-15% spd)
Avg Write: 90 MB/s (-55%)

Cost: ~240 Pounds

Summary: RAID 0+1 (F3x4) is cheaper and has twice the space, but is ~15% slower than the alternative. It offers full redundancy over it's data, so if retention of non-archived (active) work is critical, this setup has a significant edge.

SDD + RAID 1 (F3x2) has a 30% gain in speed against the primary bottleneck for Video Processing applications. The primary work data has no redundancy, increasing the risk of data loss; SSD's are much more reliable than platter disks, but there is still a small amount of data-loss risk involved, more if archiving is not performed regularly. After OS & applications are installed, working space on the drive could be anywhere from 20GB to a meager 10GB. Working by pulling data from the archive suffers significant loss: Pull data from archive at -15% speed, then process and update data at +30% speed; archive as normal during non-work hours. This comes to more time spent moving data than the improved primary drive speed saves. Redundancy on archive; data loss chances are negligible.

So what it comes down to is how much space all your applications, plug-ins, OS, and resources take on your drive, and how often you have to go into the secondary drive for resources. CS4 takes 10GB, CS5 25GB. Maya takes 4GB. 3DS Max takes 7 GB (Recommended, Including Swap). Windows 7 Professional 64-Bit takes 20 GB, +~18 GB pagefile (dependent on memory size [1.5*12GB]). Sizes of resources (surfaces etc.) on primary drive vary.

Lowest space requirements, not counting resources are 46 GB (30% Pagefile, 76%).
Highest space requirements, not counting resources are 74 GB (100%+).

You can see the dilemma with SSD here. They make good primary drives for speed, but their size is pathetic by modern standards. Even with the smallest set of applications, you only have 15 GB's for working on your drive; you'll have to run archives at LEAST daily, more likely twice a day.

Food for thought. You CAN get faster render times with the SSD, but is the frequent archiving worth the speed gain? Depends on how big your projects are. Going to CS5 later down the line is out of the question on this SSD.

Also, you didn't factor the cost of Win7. Assuming you don't already have a copy, add 90 pounds to the total cost (Win 7 Pro 64-Bit OEM System Builders' Copy)
 
Last edited:
REC, I think you need to do some reading on SSDs. Anand's SSD Relapse is a good place to start. You don't buy an SSD for the sequential read and write speeds. You buy an SSD for the random access speeds which are orders of magnitude greater than those of traditional SSDs.

With that in mind, of course you don't put project files on the SSD, that's just silly. You use it for your OS, apps, and pagefile. The point of an SSD is not to speed up render times (which it won't by any reasonable margin). The point of an SSD is to reduce the amount of time that you're waiting on the machine while you're actively editing the project.

EDIT: Oh, and where did you get those benchmarks for the F3 in RAID? They seem awfully optimistic to me.
 
Last edited:
The point of this build is to cut down render time, right? Pull resources, process data, write results. I don't dispute that SSD's have come a long way since their inception, but are the speed gains from running your OS/App's on your SSD worth it when you'll be writing the results at a meager 90 MB/s instead of 200 MB/s?

I don't know the applications' internal workings well enough to tell you, but my numbers are based on Average Read/Write speeds; i.e., a combination of both random and sequential writes, on average. The point of my post was to show that working with your files on RAID 1 would be excessively slow for the OP's needs, which was chosen in order to work with a SSD Primary Drive.

So, I yield to your opinion: Is the meager gain worth the loss? Take a look at the information, crunch some numbers, and give us a comparison.


(Note: I was unable to find anything last night about hardware render improvements because my roommate didn't pay the internet bill for 19 fscking days. I'm a HW guy more than a SW guy, so I yield to your opinion. Water under the bridge anyway, since the CPU is a beast and the vid card is pretty outstanding. 200 vs 300 USD to go up to a GTX 285, but it's out of his budget.)
 
Well the idea is that total project time = editing time + render time. SSDs won't help (much) with render times. I was mentioning that they do greatly reduce the editing time because your disk is not constantly thrashing between the project files, pagefile, and application files. While rendering, you're going to be CPU-limited, not disk limited (unless your disk is quite a lot slower than a single F3).
 
Last edited:
Keep in mind that a full length video project can be anywhere from 700 MB to 13GB, depending on resolution and compression.
 
I agree that a fast primary drive will make his application more responsive during editing, but he requested faster rendering time.

To the OP: How long are your renders? If we're talking hours on end (Some 3D renders can take literally days of processing time), I think editing time takes a back seat to rendering time.
 
REC said:
Keep in mind that a full length video project can be anywhere from 700 MB to 13GB, depending on resolution and compression.
Click to expand...

And up if we're talking uncompressed 4K. I agree that we need more info from the OP.
 
Of course, if he could afford to blow another 100 pounds he could have both, and well, the machine would be rediculous. Assuming of course he can fit all of his software on that drive (remember the 46GB quote I said was with only 30% of the recommended paging file size)
 
mfenn said:
While rendering, you're going to be CPU-limited, not disk limited (unless your disk is quite a lot slower than a single F3).
Click to expand...

The CPU isn't part of the equation at the moment, is it? We're talking the benefits of slightly faster transfer of application and OS data, versus significantly faster (75-100%) transfer of very large resource files to RAM/CPU.

In both scenarios, the CPU is static; in this case it's an i7-930.
 
REC said:
The CPU isn't part of the equation at the moment, is it? We're talking the benefits of slightly faster transfer of application and OS data, versus significantly faster (75-100&#37😉 transfer of very large resource files to RAM/CPU.

In both scenarios, the CPU is static; in this case it's an i7-930.
Click to expand...

Say what?

The application isn't going to try to buffer all of the project files in memory before it starts to work on them (and if it does, the pure read time is going to be insignificant as a % of the total time). It's going to read as it needs the data (with perhaps a tuned readahead). Who cares if you can read the raw data at 400fps vs. 200fps when the CPU can only encode at 50fps?
 
I don't mean all the resources to memory at one given time, but multiple instances of pieces of resources to memory, repeated until the resource as a whole has been processed.

Here's the stack as I see it.

Load Piece to Memory (Time per sector of read)
Process Data (Time per sector of process)
Write Piece to HDD (Time per sector of write)

Repeat above 1000 times for render.

See what I'm getting at? The speed the HDD can read and write is a factor, multiplied by the amount of times each piece must be loaded. The amount of repeats of the stack is based on the amount of frames, which controls the size of the file.
 
Last edited:
Info from the OP 🙂 :

My render times are quite decent at this moment since I own a Q9300 system with a pretty good mobo + vid card and 4GB of memory, but I`m in a position where I can afford a much faster rig..so why not go for it. Plus I`m starting university study on animation so I want a PC that "won`t stand in my way" when it comes to 3D work.

Render times are essential for me, but with a CPU like this (i7-930) it won`t be an issue. My main problem is that whenever I start to work on a bigger project, my computer starts lagging like hell. By lagging I mean: the HDD starts to work like mad, I have to wait a lot for the PC to respond every now and then. This reason behind this are the textures. JPEGS or PSD files I use during my work. If I add 7 or 8 jpegs with the size of 5000x5000 pixels..I start to feel the lagg, the torture my PC begins to suffer.

I feel that this lagg is mostly because of my poor HDD performance and maybe because the 4GB memory isn`t quite enough. Currently I`m working on a still image which has about 15 textures with resolutions above 4000x4000 pixel and the work is becomming really annoying because of this lag.

I`d like to erase this problem as much as possible.
 
bigboym said:
Info from the OP 🙂 :

My render times are quite decent at this moment since I own a Q9300 system with a pretty good mobo + vid card and 4GB of memory, but I`m in a position where I can afford a much faster rig..so why not go for it. Plus I`m starting university study on animation so I want a PC that "won`t stand in my way" when it comes to 3D work.

Render times are essential for me, but with a CPU like this (i7-930) it won`t be an issue. My main problem is that whenever I start to work on a bigger project, my computer starts lagging like hell. By lagging I mean: the HDD starts to work like mad, I have to wait a lot for the PC to respond every now and then. This reason behind this are the textures. JPEGS or PSD files I use during my work. If I add 7 or 8 jpegs with the size of 5000x5000 pixels..I start to feel the lagg, the torture my PC begins to suffer.

I feel that this lagg is mostly because of my poor HDD performance and maybe because the 4GB memory isn`t quite enough. Currently I`m working on a still image which has about 15 textures with resolutions above 4000x4000 pixel and the work is becomming really annoying because of this lag.

I`d like to erase this problem as much as possible.
Click to expand...

So your problem is not actually rendering the project, but in the editing phase? Your problem probably results from either (a) not enough memory and thus swapping or (b) trying to do too many concurrent accesses on the same disk.

If that is the case, check out an SSD to separate the application and pagefile accesses out from the other accesses and the more RAM in a new build will also help.
 
Last edited:
REC said:
I don't mean all the resources to memory at one given time, but multiple instances of pieces of resources to memory, repeated until the resource as a whole has been processed.

Here's the stack as I see it.

Load Piece to Memory (Time per sector of read)
Process Data (Time per sector of process)
Write Piece to HDD (Time per sector of write)

Repeat above 1000 times for render.

See what I'm getting at? The speed the HDD can read and write is a factor, multiplied by the amount of times each piece must be loaded. The amount of repeats of the stack is based on the amount of frames, which controls the size of the file.
Click to expand...

Right, but the the processing time is also multiplied by 1000. That time is also much greater for a typical codec than the time it takes the HDD to read the sector.
 
Working phase actually..its different from editing. Its when you work in your 3D program in your viewports. Moving stuff around, modeling, texturing etc. 12GB of ram shoudl be enough for this I guess.
 
bigboym said:
Working phase actually..its different from editing. Its when you work in your 3D program in your viewports. Moving stuff around, modeling, texturing etc. 12GB of ram shoudl be enough for this I guess.
Click to expand...

Yeah that's what I was referring to. I was saying that editing is basically everything but rendering. My point still stands though.
 
bigboym said:
...the HDD starts to work like mad...
Click to expand...

Well, I see two scenarios here. Forget the debate about processing time, let's look at application response with large resources.

When Windows starts to page, your performance gets raped. This happens when you run out of RAM; the OS starts using your Hard Drive as an extension of your memory, which works MUCH slower than RAM does. If the lag is just when you add (and thereby load) the texture, it's the Hard Drive's speed that's an issue. I don't think that's the case.

It sounds to me like you're paging during the editing portion; the solution is more memory. 12 GB's is a hell of a lot, I would hope it'd be enough but I'm just ballparking here. Windows 7 32-Bit can only support 4GB of memory, so again be sure to get 64-Bit or the extra 8 GB of memory will be useless. If you DO end up paging with 12 GB's, as mfenn pointed out to me earlier, SSD drives respond much better than HDD drives while paging.

As long as you can fit all of your software onto the 60GB SSD, it's the better bet. With everything you have though, I think you'd need an 80GB

Edit: I did a little bit of research and a 5000x5000 (25 Megapixel) .jpg texture is about 100 MB; 7-8 25 Megapixel jpg's comes to just under 1GB of RAM used when loaded, or 20% of your current RAM just loading those images. Sounds more and more like paging.
 
Last edited:
I`m paging when I start to add textures to my work. If I use low res ones.. its ok. But when I work on a really big scene (I`m talking about images with 3000x3000 res or a bit more) which requires high-res textures then I`m pretty much in hell.

I use Win7 64bit of course, 64bit software all the way because of memory. The 64GB capacity with the SSD might be an issue. I could save my work on a different drive instead, I`ll leave only the main programs (Maya, Max, ZBRush etc.) on the SSD... still, a bit more capacity would be very usefull.

So, should I stick with the rig I posted?
 
No, that's what I'm saying; have your OS/Applications on an SSD and have your data on another drive (or RAID), but a 64 GB SSD will be too small. Just your OS and applications will easily consume 50 GB of space, probably 60 GB because of the extra large page-file required for 12 GB of RAM (18 GB Pagefile)
 
Last edited:
bigboym said:
I`m paging when I start to add textures to my work. If I use low res ones.. its ok. But when I work on a really big scene (I`m talking about images with 3000x3000 res or a bit more) which requires high-res textures then I`m pretty much in hell.

I use Win7 64bit of course, 64bit software all the way because of memory. The 64GB capacity with the SSD might be an issue. I could save my work on a different drive instead, I`ll leave only the main programs (Maya, Max, ZBRush etc.) on the SSD... still, a bit more capacity would be very usefull.

So, should I stick with the rig I posted?
Click to expand...

I think the rig you posted is very good and that you should go with that. Don't worry about trying to keep all of your data on the SSD, just the OS and main programs.
 
Okay guys, thanks very much for your help. I`ll calculate the space requried for main programs and if they fit I`ll stick with an SSD. If not, I`ll ditch it.

Thanks again!
 
Hello,

I thought I had my final setup for my new PC, but some information has been brought to my attention which indicates that an i7 875K + a good P55 chipset motherboard would be a better choice instead of the original, i7 930 + ASUS P6TD Deluxe.

A friend of mine says that the high end P55 motherboards have memory controllers which are faster than the one in the CPU. The max speed on the 930 with a X85 mobo is 25.6gb/s which the cpu can work with the memory. Although the P55's memory controller with a... lets say 8GB DDR3 2000 memory in dual mode will have 2x16GB/s speed which is 32GB/s.. much more that the 25.6GB/s.

I`m kind of lost here, and the info might not be very clear. I`d like to sort this out.
This person says that I can run DDR 1600 or even 2000 memory much faster on a P55 platform than on a X58 one. This is because of the memory controller which is on the P55 mobo. The X58 doesn`t have such a memory controller.

I`d appriciate if someone could shed some light on this matter.
Thank you!
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top