Ok. Yes I have searched this forum and elsewhere and discovered that NVidia cards are the best for Premiere Pro CS5. My budget now will only allow up to $200 or maybe $300 if it's really worth it. Is there any Nvidia cards that fit into this price range that will perform for CS5?
Many thanks in advance for all your replies.
Harry

nVidia GTX-460

Zotac GTX 460 2GB with DVI, HDMI and DisplayPort: techPowerUp! News :: Zotac Readies Non-Reference GTX 460 with 2 GB Memory (http://www.techpowerup.com/127813/Zotac_Readies_Non-Reference_GTX_460_with_2_GB_Memory.html)

"No word on availability or pricing." My guess is that non-reference card will be well above the OP's budget.

Many Thanks everyone for you replies. That's what I was looking for.
Harry

nVidia GTX-460

Harm:
Will this card work without any hacking? In other words right out of the box?
Thanks again for your help.
Harry

You will need to do the hack.

As David stated, you will need to do the "hack" (which, in this case, is tinkering with a couple of configuration files).

As for the currently produced GPUs, all of the ones that are officially supported without the "hack" are either EOL or about to be EOL'd. The Fermi-based Quadros will replace the current Quadros.

I went to Newegg.com and found several NVIDEA cards. Need alittle help on which to chose.
Thanks Again

Newegg.com - Computer Hardware,Video Cards & Video Devices,Desktop Graphics / Video Cards,GeForce GTX 400 series,GeForce GTX 460 (Fermi) (http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=100007709%20600007323%20600062521&IsNodeId=1&Tpk=GeForce%20GTX%20460)

Limit your choice to cards that have at least 1 GB of memory.

I agree with Harm, stay with the 1 GB memory cards. Also, I recommend going with an overclock card that has dual cooling fans. This will help to keep the video card cool so you don't have any perfomance issues with the Mercury Playback Engine.

Limit your choice to cards that have at least 1 GB of memory.

I agree with this, especially since most of the 1GB GTX 460 cards cost just a few dollars more than their 768MB brandmates (if they are in stock). Plus, the 1GB cards have a 256-bit GDDR5 memory bus versus only a 192-bit bus on the 768MB versions.

Add to this that Adobe requires 896 or something like 860+ MB for the hack to work. 768 MB does not allow the hack to work. So better be safe than sorry and go for 1 MB+ memory.

Harm--The GPUSniffer app actually says it needs 765MB, so a 768 might work--here's a sample from my laptop:

"CUDA Device # 0 not choosen because 765MB are required, and 111MB are present."


And to the OP, we just purchased this card: Newegg.com - GIGABYTE GV-N460OC-1GI GeForce GTX 460 (Fermi) 1GB 256-bit GDDR5 PCI Express 2.0 x16 HDCP Ready SLI Support Video Card (http://www.newegg.com/Product/Product.aspx?Item=N82E16814125333)

It has two fans which are inaudible from outside the case, even after 30 mins of CUDA enabled encoding in Premiere.

Harm--The GPUSniffer app actually says it needs 765MB, so a 768 might work--here's a sample from my laptop:

"CUDA Device # 0 not choosen because 765MB are required, and 111MB are present."

Actually, that 765MB requirement refers to unused (free) graphics RAM, not total graphics RAM. And a 768MB card will not work with MPE's GPU acceleration feature because at Adobe's minimum required resolution of 1280x900, more than 4MB of graphics RAM would have already been used up with the desktop color depth is set at 32 bpp (High color). And more graphics RAM than that would already be taken up by Windows Aero and Windows graphical enhancements that are included in the OS itself. (And those amounts of used graphics RAM put together are from just idling at the Windows desktop!) Thus, by the time CS5 loads up, nearly 20MB of that 768MB of RAM would have already been used up by Windows itself.

Randall is 100% correct on the ram requirement. Someone with a 768MB card learned this early on in the 'hack' thread.

Something else to consider is the warranty. Some have a few years and some have lifetime (PNY & EVGA, I think).

Harm--The GPUSniffer app actually says it needs 765MB, so a 768 might work--here's a sample from my laptop:

"CUDA Device # 0 not choosen because 765MB are required, and 111MB are present."


And to the OP, we just purchased this card: Newegg.com - GIGABYTE GV-N460OC-1GI GeForce GTX 460 (Fermi) 1GB 256-bit GDDR5 PCI Express 2.0 x16 HDCP Ready SLI Support Video Card (http://www.newegg.com/Product/Product.aspx?Item=N82E16814125333)

It has two fans which are inaudible from outside the case, even after 30 mins of CUDA enabled encoding in Premiere.

Hi Everyone:
I just ordered the above Card from Newegg.com. I was impressed with what I read. Plus the fact that it's in my price range. I report back after I install it and put it to use.
Thanks again for all your replies.
Harry

Something else to consider is the warranty. Some have a few years and some have lifetime (PNY & EVGA, I think).

And with both of those latter brands (eVGA and PNY), only selected cards have a lifetime warranty - namely, the eVGA cards that have their full model numbers ending in "-AR" and those PNY XLR8 line of cards. The other eVGA cards have only a two-year warranty while PNY's lower-level Verto line have a three-year warranty.

To the OP (Harry):

Excellent call on the 1GB GTX 460. However, the few results currently in the PPBM5 list on their Web site for the GTX 460 is slower than most other MPE-enabled cards because the currently available drivers for that card are not mature enough.

To everybody who responded to this thread:

IMHO a GTX 260 or higher is the best choice for CS5. Lower-class cards with 1GB or more RAM are not as good, but still provide a significant boost. The worst choice for CS5 are those NVIDIA cards which cannot use MPE's GPU acceleration mode (either due to the lack of CUDA or an insufficient amount of graphics RAM) because they perform even slower than ATi cards in software-only mode.

Long time lurker, first time poster.

Currently have the Gigabyte 460 1GB model. Able to get it on sale for $199 recently. Not too loud at all.
Seems to be the best value per performance card that nvidia has produced in a while. Works great with MPE and the hack. Overall, probably the best card performance for price that I have owned in a long time.

Currently MPE is still in it's infancy and CUDA Toolkit 3.1 has just appeared.

What does that mean? First of all that at this moment in time there is no performance difference noticeable between a GTX-285 and a GTX-480, tested on the same system. This despite more than double the cores and double the width of the memory bus of the 480 over the 285.

At this moment the conclusion may be that the 460 is the most attractively priced MPE card with the same performance as the 480.

However, it is to be expected that when MPE get's more mature, it will use available cores more efficiently. Just like the threading on certain processes in CS5 that is far from optimal, causing much more latency in the communication between CPU - RAM - GPU - VRAM and back, especially on hexa cores.

Threading is a major factor for performance gains, both in the CPU and in the GPU. Therefore I expect that in time the 470 or 480 will be a much better choice than the 460.

Harm,

I hope you are right about future optimizations improving the performance of these video cards, but do the 470/480 offer any other meaningful architectural advantages over the 460 beyond the "processor core" count? I know the memory bandwidth on the 480 is much higher Do the benchmarks show memory size/bandwidth (beyond the previously discussed limit) really matter for performance?

FWIW, here's a little info I gleaned off a simple search on Newegg:

GTX 260 896MB: 216 cores $180 $0.83 per core
GTX 285 1 GB: 240 cores $300 $1.25 per core
GTX 460 1 GB: 336 cores $230 $0.68 per core
GTX 470 1.2GB: 448 cores $350 $0.78 per core
GTX 480 1.5GB: 480 cores $450 $0.94 per core

To everybody who responded to this thread:

IMHO a GTX 260 or higher is the best choice for CS5. Lower-class cards with 1GB or more RAM are not as good, but still provide a significant boost. The worst choice for CS5 are those NVIDIA cards which cannot use MPE's GPU acceleration mode (either due to the lack of CUDA or an insufficient amount of graphics RAM) because they perform even slower than ATi cards in software-only mode.

To the first two sentences in the above quote, it also depends on the CPU used. After all, it is way overkill to put in a GTX 480 in a Core 2 Duo system because the GPU would have been significantly faster than the CPU/system RAM subsystem. If CS5 were to run on such an imbalanced system, the GPU would then have to wait for the CPU to catch up.

On the other hand, some of the low-end GPUs have too slow of a graphics memory subsystem (read: too low of a memory bandwidth) to be of any good in MPE. Those cards which use 1GB of DDR2 memory showed no performance improvement (or even a performance drop) versus software-only mode. Hence, those bottom-of-the-line cards are not recommended for CS5 even if they have 2GB of DDR2 graphics RAM.

As for the last sentence, it all boils down to the different optimizations in the different manufacturers' drivers. In my testing, ATi's drivers are actually significantly faster than NVIDIA's drivers in MPEG-2 SD encoding performance but moderately slower in H.264 HD encoding performance.

Randall,

Just today I have updated both the Interpreting Results page and the Background Information page, to give some more information about the differences between hardware and software MPE when encoding. It may help you understand what causes the differences between ATI and nVidia cards and MPE on/off results.

However, keep in mind the maximum quality settings when using CUDA/MPE.

PPBM5 Benchmark (http://ppbm5.com/Index.html)

Harm,

Can you point me to where it says that Maximum Quality is engaged when using hardware acceleration because MRQ is not engaged by default or in PPBM5. Also, using MRQ greatly increases quality as well as rendering time, but it is worth every extra second of rendering time.

So I went with the 480 in hopes there will be future enhancements down the road. I swapped the new 480 into my Supermicro Xeon system and re-ran the PPBM5 rendering test only. I did not run the encoding tests as I assumed these would be unaffected by the new card.

With MPE SW acceleration only the rendering time was 199 secs, virtually the same as with my older ATI card.

With GPU Acceleration turned on after the hack, the rendering time was 20 secs – a tenfold increase. Remarkable.

Good God, that 480 is a huge beast, with what looks like chrome exhaust pipes wrapping around the body and a fan that extends over the CPU heat sinks (or something – it barely fits into my huge Supermicro case). It takes THREE 6-pin power supply cables, as the only unused 8-pin connector from my PSU was the wrong type. Fortunately EVGA included an 8-pin to Two 6-pin adapter cable, and all fits and seems to be working well. At least there is no apparent smoke that I can see or smell.

By the way, a simpler form of the hack is to simply find the txt file as directed, and replace the “285” in “GeForce GTX 285” with “XXX”, where XXX is your model number. Reduces the chance of typos this way and eliminates the need to go into that other exe file and all that nonsense. I suppose you could do the same with the Quadro card numbers as well. We observed from other posts that even a capitalization error renders the hack ineffective.

So I went with the 480 in hopes there will be future enhancements down the road. I swapped the new 480 into my Supermicro Xeon system and re-ran the PPBM5 rendering test only. I did not run the encoding tests as I assumed these would be unaffected by the new card.

With MPE SW acceleration only the rendering time was 199 secs, virtually the same as with my older ATI card.

With GPU Acceleration turned on after the hack, the rendering time was 20 secs – a tenfold increase. Remarkable.

In my experience with the GTX 470 (versus the HD 5770) the MPEG-2 encoding speed was actually slower with the 470 than with the 5770 but the H.264 encodes were faster. Also, with MPE SW acceleration only the 5770 was slightly slower in timeline rendering speed than the 470.

Also, I found differences between the 257.21 and the 258.96 drivers in my testing with the 470. The 257.21 drivers were a bit faster in encoding but the encodes suffered in quality, with black or green squares ruining the encodes through the 257.21s.

And of course the difference between MPE in software-only mode and MPE in GPU-accelerated mode with the 470 is roughly ten-fold.

But while the 470 is a great card for PPro CS5 use, it is not really needed there (though a system with a highly overclocked hexacore CPU would have benefitted from the 470 over the GTX 2## series GPUs). I could have gotten nearly the same level of performance by going as low as a 1GB GTS 250 (at least with my particular system's CPU). And had I continued to use my secondary system's Core 2 Quad Q9450 for CS5, I could have gotten away with even a 1GB DDR3 or GDDR5 version of the GT 240.

I posted this in another thread and thought it would add to the discussion here:

Here's an excerpt from David Knarr's well-written article:

To give you an example of the performance difference between DDR3 and DDR5 memory, we ran some tests using a GT240 video card with 1 gig of DDR5 memory and a GT240 video card with 1 gig of DDR3. The video card with the DDR5 memory gave us around a 45% speed increase over a GT240 with DDR3 memory.

We also tested the GT240 with DDR5 memory against a GT250, GT260 and a GT 430, with DDR3 memory, The GT240 with DDR5 memory was faster in every test, even though some of the video cards had more CUDA cores. This was due to the GT240 having DDR5 memory and the fact that the Mercury Playback Engine isn't using all of the CUDA cores.

We are trying to find out exactly how many CUDA core the Mercury Playback Engine is using. I have heard rumors that it only use around 100 CUDA cores. Now remember, this is only a rumor. But from the tests we have run and what a few other people are reporting, I would have to say this may be true.


What I get from this is that the most important criteria in choosing a card for MPE playback right now (not taking into consideration potential in the future) are: 1) DDR5 memory, and 2) at least 100 cores (96 cores is close enough).

A logical conclusion, if these statements are true, is that having more cores only provides marginal improvements.

Elsewhere in his article, he says:

Right now, you will get the same performance across all the GTX 400 series of video cards. Right now, the performance is about the same as the GTX240 to GTX 285 video cards.


It would be great if you could chime in, David and Harm, with your experiences and if there are any considerations for buying a more expensive card than a GT240 with DDR5 memory right now or in the very near future.

It would be very exciting if a GT240 provides close to the same performance as a GTX480 when it comes to Mercury Playback. I don't play games and am just interested in Premiere CS5 performance. Also, I believe the GT2xx series is compatible with OSX, whereas there are no Fermi (4xx series) drivers (yet) for OSX.

Thanks in advance!

I posted this in another thread and thought it would add to the discussion here:

Here's an excerpt from David Knarr's well-written article:


What I get from this is that the most important criteria in choosing a card for MPE playback right now (not taking into consideration potential in the future) are: 1) DDR5 memory, and 2) at least 100 cores (96 cores is close enough).

A logical conclusion, if these statements are true, is that having more cores only provides marginal improvements.

Elsewhere in his article, he says:


It would be great if you could chime in, David and Harm, with your experiences and if there are any considerations for buying a more expensive card than a GT240 with DDR5 memory right now or in the very near future.

It would be very exciting if a GT240 provides close to the same performance as a GTX480 when it comes to Mercury Playback. I don't play games and am just interested in Premiere CS5 performance. Also, I believe the GT2xx series is compatible with OSX, whereas there are no Fermi (4xx series) drivers (yet) for OSX.

Thanks in advance!

The 1GB DDR5 version of the GT 240 or the newer GTS 450 would at least suffice for most systems, especially the "consumer" systems that are to be used for CS5. The only way that the 1GB DDR5 version of the GT 240 might be limited would be if you're stuck with an early Intel Core 2 Duo or an AMD CPU with fewer than four cores.

But what about the ram's Bit? Both of the GT 240s only use 128-bit whereas other cards go up to 448-bit with DDR3 ram, and this increases bandwidth to similar, if not, greater than DDR5/128-bit.

But what about the ram's Bit? Both of the GT 240s only use 128-bit whereas other cards go up to 448-bit with DDR3 ram, and this increases bandwidth to similar, if not, greater than DDR5/128-bit.

This is because in my testing MPE barely takes full advantage of even 128-bit graphics memory, let alone 448-bit.

The 1GB DDR5 version of the GT 240 or the newer GTS 450 would at least suffice for most systems, especially the "consumer" systems that are to be used for CS5. The only way that the 1GB DDR5 version of the GT 240 might be limited would be if you're stuck with an early Intel Core 2 Duo or an AMD CPU with fewer than four cores.

Randall, We have Premiere CS5 running on a couple of AMD Athlon 64 5200+ systems. Both of these systems have a clock speed of 2.7Ghz and have 6 gigs of memory. These are both HP a1630n stock computers and we are running the GT240 video cards in them. They do okay with Premiere CS5. The GT240 gave us a nice improvement. So I would recommend the GT240 for the lower end systems to.

It would be great if you could chime in, David and Harm, with your experiences and if there are any considerations for buying a more expensive card than a GT240 with DDR5 memory right now or in the very near future.

It would be very exciting if a GT240 provides close to the same performance as a GTX480 when it comes to Mercury Playback. I don't play games and am just interested in Premiere CS5 performance. Also, I believe the GT2xx series is compatible with OSX, whereas there are no Fermi (4xx series) drivers (yet) for OSX.

Thanks in advance!

Peter, the GT240 on the systems I used here to test with (an AMD X4 and AMD X6) I am getting the same results as with the 450. Now keep in mind, if and when Adobe opens up the Mercury Playback Engine to take advantage of more CUDA cores, the the 450 would be faster. Also, I have not tested these video cards on anything other than AMD systems.

I don't personnaly have access to any systems with the I7 CPU. But these are my results so far on the AMD CPUs.

Just an update, I am not advocating getting the GT240, but right now it is performing as well as the higher end video cards in our systems.

Thanks, David. Are there any reasons why you are not advocating getting the GT240? Is there a card you do advocate?

Just an update, I am not advocating getting the GT240, but right now it is performing as well as the higher end video cards in our systems.

Peter, the only reason why I said that, was everyone has their favorite card (ie. GT285, GT450, etc), I just didn't want to promote one card over another.

I have 6 of the GT240s installed, along with a GT260 and GT460 installed in our systems. Personally, for the money and currenty performance in Premiere CS5 I like the GT240 with DDR5 memory. In our systems, it is performing the same as the other cards with Premiere.

BTW, you mentioned the article with all updates happening often. I am trying to figure out a way to make the updates to our article standout, like putting the updates in red. The only problem has been, is I have been updating every few days and within a week or two, alot of it will be in red.... lol.

BTW, you mentioned the article with all updates happening often. I am trying to figure out a way to make the updates to our article standout, like putting the updates in red. The only problem has been, is I have been updating every few days and within a week or two, alot of it will be in red.... lol.

Maybe use a different color for each new update or a little blurb about what section(s) were updated? It was hard for me to look through and see what changed. Anyways, I appreciate all the information and experience you share in your article.

Randall, that makes sense as to why they all perform similarly.

If you look on the ppbm5.com site, there is one user, BillG, that seems to have tested his i7 980x setup with different cards. His setup comparing the 285 vs 480 yields almost identical results. It looks like he has boosted his performance gains even more by overclocking his CPU. He sped up the benchmark by 50 seconds just from overclocking from 4.2 to 4.4GHz, if I am interpreting the results correctly. Very interesting!

Harm, is there any way to download the spreadsheet data so that we can sort the results to draw better conclusions?

Randall, that makes sense as to why they all perform similarly.

If you look on the ppbm5.com site, there is one user, BillG, that seems to have tested his i7 980x setup with different cards. His setup comparing the 285 vs 480 yields almost identical results.

That despite the 285 having a 512-bit DDR3 bus versus the 384-bit DDR5 bus on the 480 (and the 480 having twice as many cores as the 285).

Peter,

Bill and I have been looking into that, but we encounter several problems. A SQL database would offer the sortability you so much desire (and you are not unique in that) but loses the colors and conditional formatting (Top, D9, Q3, Med, etc.) so in our view that is out. Apart from the ASP and NET efforts from our side. Keep in mind that both Bill and I do this as a hobby and a service to you all.

We have been looking into publishing pre-sorted result pages for the top-xx submissions, that are static by nature. The drawback for us is the effort it takes to update all these sheets every time we get another top-xx result. xx being undefined at the moment.

We have been looking into adding a down-loadable version of the spreadsheet, but we do want to protect the macros, conditional formatting and formulas used in the spreadsheet, so that means we have to create a new spreadsheet with values only and that creates new problems.

Rest assured, we are looking into this, but have not yet found a sensible solution. As soon as we have found one, we will let you know on the PPBM5 site.

That despite the 285 having a 512-bit DDR3 bus versus the 384-bit DDR5 bus on the 480 (and the 480 having twice as many cores as the 285).

I would think the 285 would be much slower due to the DDR3 memory...

Maybe use a different color for each new update or a little blurb about what section(s) were updated? It was hard for me to look through and see what changed. Anyways, I appreciate all the information and experience you share in your article.

Peter I have tried using different colors offline and after about a month, it starts to look really bad. I think what I am going to do is at the end of the article I have put a section with the date and the update made to the main article.

Right now the article is pretty much complete with the exception of new driver updates and new video cards to add to the list, so I think my idea of putting the update info at the end would work. At least I hope so.... lol.

Peter,

Bill and I have been looking into that, but we encounter several problems. A SQL database would offer the sortability you so much desire (and you are not unique in that) but loses the colors and conditional formatting (Top, D9, Q3, Med, etc.) so in our view that is out. Apart from the ASP and NET efforts from our side. Keep in mind that both Bill and I do this as a hobby and a service to you all.

We have been looking into publishing pre-sorted result pages for the top-xx submissions, that are static by nature. The drawback for us is the effort it takes to update all these sheets every time we get another top-xx result. xx being undefined at the moment.

We have been looking into adding a down-loadable version of the spreadsheet, but we do want to protect the macros, conditional formatting and formulas used in the spreadsheet, so that means we have to create a new spreadsheet with values only and that creates new problems.

Rest assured, we are looking into this, but have not yet found a sensible solution. As soon as we have found one, we will let you know on the PPBM5 site.

Thank you, Harm and Bill, for your service. I am trying to get some insight into optimizing for CS5 performance thanks to your benchmarks.

I was having problems trying to just copy and paste from the benchmark results page into excel as it would be tab-deliminated data and wouldn't import properly into excel. However, I was able to get the data into Excel via a web query.

Just being able to sort through and filter things makes it easier to analyze the data and draw conclusions.

Anyways, I already appreciate all the work you've done putting the tests and site together and updating them.

Thanks again,
Peter

I noticed that some of the faster Disk Access setups also have faster non-MPE speeds. Most of the faster setups have RAID0 setups, some with SSD drives and some with up to an 8 drive RAID0! Is this even safe? Wouldn't it be disastrous if even 1 drive were to go bad, especially in BillG's 8 drive RAID0 setup?

If even one drive in a RAID 0 setup goes bad, the data (if any) on the other seven becomes unusable. (At least without the need to use astronomically expensive data recovery services.)

Though I have to admit that even single-drive setups are not immune to total data loss. In fact, the very slowest Disk Access setups are those systems that use just a single drive (or single drive volume) for everything - the OS, programs, projects, renders, page file and video source and output files. Those are the systems that have the greatest risk for data loss due to the extra wear and tear on the single drive. Those are the very reasons why Adobe does not recommend single-drive systems at all with any of its prosumer-level Creative Suite applications.

And yes, a two-drive software RAID 0 array on an Intel SATA controller does not improve the disk access scores enough over a very fast single disk (separate from the OS disk) to justify the trouble of creating such an array.

Peter,

This is a very valid and good point you make. Normally no one would even consider a 8 disk raid0 for editing, but for benchmarking it is OK, at least for that purpose only, IMO.

Rendering does require that the preview files are written to disk and then disk speed does come into the picture as well, not only CPU/GPU/memory speed.

Bill is constantly trying out various disk setups, with multi SSD's, SCSI's, SATA's with different cards (LSI, ARECA 1680, and shortly 1880) and different raid configurations. I always wonder how his basement looks with all the stuff he gathered in the past years.

It does make for interesting scenarios just to see "what if?" However, as you mention, it is not safe in a real world application to have crazy RAID0 setups. I can't imagine trying to back up your 8 TB setup in case you do get a disk failure...

I think performance can also be tweaked by specifying the speed and latency of the RAM used. From articles I have read, the latency improves performance more than the speed of the RAM. I wonder if this has any effect on the benchmarks since we know the amount of RAM available definitely makes a difference.

It does make for interesting scenarios just to see "what if?" However, as you mention, it is not safe in a real world application to have crazy RAID0 setups. I can't imagine trying to back up your 8 TB setup in case you do get a disk failure...

I think performance can also be tweaked by specifying the speed and latency of the RAM used. From articles I have read, the latency improves performance more than the speed of the RAM. I wonder if this has any effect on the benchmarks since we know the amount of RAM available definitely makes a difference.

This is not the case, based on my testing. In fact, memory latency ratings have very little impact on the CS5 performance. What's more, CL6 or CL7 memory @ DDR3-1600 speed might actually perform slower than CL9 or CL10 memory at the same memory speed unless there is crazy cooling on the CPU because such ultra-low-latency memory typically requires a much higher DIMM voltage than standard DDR3 memory at the same memory clock speed. And this is all because raising the DIMM voltage also increases the voltage of the memory controller, which in turn requires a raising of all of the other CPU voltages as well in order to minimize damage - and all of that raises the temperature of the CPU itself significantly. With some CL6 or CL7 1600-speed memory and weaker IMCs, the CPU might start throttling back at even stock CPU speeds. That might have forced the user to severely underclock the CPU in order to keep temps under control.

And yes, a two-drive software RAID 0 array on an Intel SATA controller does not improve the disk access scores enough over a very fast single disk (separate from the OS disk) to justify the trouble of creating such an array.

Really? I thought a RAID0 gives 2x the speed? The only benchmarks I could compare were from Dieter's_PC and Dieter's_PC_upd. He changed to a RAID0 and his MPEG2-DVD was almost a minute faster from 153s down to 98s.

This is not the case, based on my testing. In fact, memory latency ratings have very little impact on the CS5 performance. What's more, CL6 or CL7 memory @ DDR3-1600 speed might actually perform slower than CL9 or CL10 memory at the same memory speed unless there is crazy cooling on the CPU because such ultra-low-latency memory typically requires a much higher DIMM voltage than standard DDR3 memory at the same memory clock speed. And this is all because raising the DIMM voltage also increases the voltage of the memory controller, which in turn requires a raising of all of the other CPU voltages as well in order to minimize damage - and all of that raises the temperature of the CPU itself significantly. With some CL6 or CL7 1600-speed memory and weaker IMCs, the CPU might start throttling back at even stock CPU speeds. That might have forced the user to severely underclock the CPU in order to keep temps under control.

So, just get the cheapest RAM available? That would be great news :)

Peter,

Keep in mind that there is a difference between disk access time and transfer rate.

In general creating a raid0 does hardly anything to disk access, it remains approximately the same or slightly faster, but that difference is negligent. What does change is the CPU load, especially with software raids, that increases a bit but again not enough to be worth talking about and - this is where the advantage is - it increases the transfer rate of the data almost linearly. A 2 disk raid0 is nearly twice as fast as a single disk, a three disk raid0 is nearly three times faster than a single disk, etc. Again not in disk access but in transfer rate.

A modern day disk can achieve transfer rates of around 100 MB/s, a 2 disk raid0 can achieve almost 200 MB/s but in both cases the disk access will be around 13 ms (based on 7200 RPM disks).

One caveat: Do not expect that this linear behavior continues indefinitely. The bandwidth of the PCIe bus will at some point become the bottleneck. On my array (12 disk raid30) I can't get better performance than average 853 MB/s transfer with a disk access of 10.7 ms (Areca based). A single disk comes out with about 100 MB/s transfer and 13.5 ms access and a two disk raid0 comes out with around 180 MB/s and 15.8 ms access (Marvell based). BTW this is all based on the old Samsung F1 disks and the newer F3's are a lot faster.