Need help on video editing computer build

[lloyd_joy]

---

^ check the link sa review bro if acceptable ra ba ang time difference nila... kana, official review gyud na sya... they did it with a correct test methodology... not just by feeling... but, as you've said, the GTX 560 is way faster than the GT 640... but again, in video encoding, it is only in seconds or minutes... dili man siguro mo abot ug 5 minutes ang time difference... pero, depende sa ka dak-on sa file siguro... got no idea gyud pud... hehehe...

i'm not a video editor man pud hinoon as money generating activity... had just based my judgement on the review as shared nko sa taas... gamit ko for video encoding diri for hobby lang. panagsa ra pud. kun ganahan lang. got a dedicated rig for this activity. para pud magamit ni kay extra rig man gud nko sya... it is just a G840 w/ 8GB ram and GTX 650 2GB VRAM with 1TB HDD... wala ra ko maka test using my GT 640 and GTX 670 kun unsa ka dugay ang time difference sa processing ani nila...

[werty]

get something with more cuda cores.

[lloyd_joy]

bro werty,

not necessary na statement nang the more cuda cores, the better... kani ai... "cut and paste" nko ni ha... based sa recommendation sa review, bang for the money would be the GTX 660 (Ti or Non-Ti variants)...

Adobe After Effects CS6 and CC Video Cards with CUDA Acceleration Ray Trace Rendering Engine Unlock Enable MPE Hack Mod Tip David Knarr



Some people think they should run out and buy a video card with the most cuda cores (for example a GTX-680). However, that may not be the best idea, especially with programs that use OpenGL, such as After Effects.

The more CUDA cores, the wider memory interface width and the faster memory bandwidth, the better performance the video card has to offer. Remember, just because the video card has more to offer, doesn't mean the rest of the computer system will take advantage of it. Most of the time the video card's GPU will be waiting on the rest of the computer to feed it the data to process. When the GPU gets a hold of the data, the more CUDA cores the faster it will process the data. Then it hands this data back to the CPU. The wider the memory interface width and the faster the memory bandwidth, the faster the data will move back and forth to the CPU.

For example, the GTX 570 video card has 480 CUDA cores, while the GTX680 has 1,536 CUDA cores. Since the GTX680 has more than 3X the number of CUDA cores over the GTX570, you might expect the GTX680 to be 3X faster than the GTX570. But, they are not. They are only few seconds difference. The reason is, the memory interface width is smaller on the GTX6XX series of video cards, even though they have more cuda cores. Where the GTX5XX series of video cards have a much wider memory interface width, which makes up for them having fewer cuda cores.

On some tests I have done with After Effects using OpenGL plugins, the GTX570 card with it's wider memory interface performed about the same as the GTX680. In this case, OpenGL doesn't care about the number of cuda cores. It mainly cares about the faster memory bandwidth and the GPU clock speed.

However, when using the Ray Trace 3D Render Engine, the GTX 680 was faster than the GTX 570.

Then GT 600 and GTX 600 series of video cards have a new design and chipset over the older 200, 400 and 500 series of video cards, that allow them to be slightly faster, especially when using the Ray Trace 3D Render Engine since it uses the Cuda cores.

The Raytrace 3D Render Engine will render faster with more Cuda Cores that the GTX600 and GTX700 series of video cards offer. The newer GTX700 series video cards seem to have a much better memory bandwidth, as they have a higher transfer rate.

In addition, the 600/700 series of video cards run cooler and require less power than the older video cards.

If you have an older computer, such as a quad core running at 2.5 Ghz with only 4 gigs of memory, then buying the GTX-680 would be over kill. On this system, you are better off adding more memory to the computer and using a GTX-660 video card.

Since everyone has a different level of acceptable editing performance. Some people are fine with a working on a slower system, while other people need a high end, super fast system. A lot depends on whether they are a hobbyist, semi-pro or a professional editor, as to what their needs are.

With After Effects there are a lot of things that make up the system performance. Having a fast CPU, with 4, 6 or 8 cores, lot's of memory 8 gigs bare minimum, with 16+ gigs being better.

I have run some tests with the Ray Trace 3D Render Engine, on several different computer systems using a GT240, GTX 650 TI, GTX 660 and a GTX 680 video cards. I found the GTX 660 giving me the best performance for the money.

With the test I ran, the GTX 680 was only a few seconds faster than the GTX 660, except when it came to using the Raytrace 3D render engine. Here the GTX-680 was much faster.

The best bang for the buck video card for After Effects seems to be the GTX 660 or GTX 660 Ti. If you have the money and want all the speed you can get, then look at the GTX Titan or GTX 780 video card. Both of these video cards are super fast with the Raytrace 3D Render Engine and Elements 3D.

As far as what video card you should buy, that is really up to you and your budget. If it was me, I would go with the GTX 660. I will be doing more tests as I have time in 2013.

[werty]

more cuda cores than the videocard he was planning to get, not necessarily the MORE it has, the better...

as i have said "get something with more cuda cores"

[Sempron Guy]

bro werty,

not necessary na statement nang the more cuda cores, the better... kani ai... "cut and paste" nko ni ha... based sa recommendation sa review, bang for the money would be the GTX 660 (Ti or Non-Ti variants)...

Adobe After Effects CS6 and CC Video Cards with CUDA Acceleration Ray Trace Rendering Engine Unlock Enable MPE Hack Mod Tip David Knarr



Some people think they should run out and buy a video card with the most cuda cores (for example a GTX-680). However, that may not be the best idea, especially with programs that use OpenGL, such as After Effects.

The more CUDA cores, the wider memory interface width and the faster memory bandwidth, the better performance the video card has to offer. Remember, just because the video card has more to offer, doesn't mean the rest of the computer system will take advantage of it. Most of the time the video card's GPU will be waiting on the rest of the computer to feed it the data to process. When the GPU gets a hold of the data, the more CUDA cores the faster it will process the data. Then it hands this data back to the CPU. The wider the memory interface width and the faster the memory bandwidth, the faster the data will move back and forth to the CPU.

For example, the GTX 570 video card has 480 CUDA cores, while the GTX680 has 1,536 CUDA cores. Since the GTX680 has more than 3X the number of CUDA cores over the GTX570, you might expect the GTX680 to be 3X faster than the GTX570. But, they are not. They are only few seconds difference. The reason is, the memory interface width is smaller on the GTX6XX series of video cards, even though they have more cuda cores. Where the GTX5XX series of video cards have a much wider memory interface width, which makes up for them having fewer cuda cores.

On some tests I have done with After Effects using OpenGL plugins, the GTX570 card with it's wider memory interface performed about the same as the GTX680. In this case, OpenGL doesn't care about the number of cuda cores. It mainly cares about the faster memory bandwidth and the GPU clock speed.

However, when using the Ray Trace 3D Render Engine, the GTX 680 was faster than the GTX 570.

Then GT 600 and GTX 600 series of video cards have a new design and chipset over the older 200, 400 and 500 series of video cards, that allow them to be slightly faster, especially when using the Ray Trace 3D Render Engine since it uses the Cuda cores.

The Raytrace 3D Render Engine will render faster with more Cuda Cores that the GTX600 and GTX700 series of video cards offer. The newer GTX700 series video cards seem to have a much better memory bandwidth, as they have a higher transfer rate.

In addition, the 600/700 series of video cards run cooler and require less power than the older video cards.

If you have an older computer, such as a quad core running at 2.5 Ghz with only 4 gigs of memory, then buying the GTX-680 would be over kill. On this system, you are better off adding more memory to the computer and using a GTX-660 video card.

Since everyone has a different level of acceptable editing performance. Some people are fine with a working on a slower system, while other people need a high end, super fast system. A lot depends on whether they are a hobbyist, semi-pro or a professional editor, as to what their needs are.

With After Effects there are a lot of things that make up the system performance. Having a fast CPU, with 4, 6 or 8 cores, lot's of memory 8 gigs bare minimum, with 16+ gigs being better.

I have run some tests with the Ray Trace 3D Render Engine, on several different computer systems using a GT240, GTX 650 TI, GTX 660 and a GTX 680 video cards. I found the GTX 660 giving me the best performance for the money.

With the test I ran, the GTX 680 was only a few seconds faster than the GTX 660, except when it came to using the Raytrace 3D render engine. Here the GTX-680 was much faster.

The best bang for the buck video card for After Effects seems to be the GTX 660 or GTX 660 Ti. If you have the money and want all the speed you can get, then look at the GTX Titan or GTX 780 video card. Both of these video cards are super fast with the Raytrace 3D Render Engine and Elements 3D.

As far as what video card you should buy, that is really up to you and your budget. If it was me, I would go with the GTX 660. I will be doing more tests as I have time in 2013.

Cause Fermi has different implementation in terms of cuda vs kepler. Fermi cuda cores =/ Kepler cuda cores. That's apples to oranges comparison ang gibuhat sa reviewer.

[lloyd_joy]

bro werty,

ah i see... if dili lang mag games si TS, kun mag focus sya sa video editing lang... that GTX 650 Ti would be just fine... no need to spend money on those more expensive GPUs w/ higher count CUDA cores... gamay ra kaayo performance difference nila... the higher CUDA cores is only a "nice to have"... naa man gud nang review sa link na akoang gi post... maka kuha ta'g idea kun worth ba ang performance to price ratio for the chosen application...

if sagulan ug GPU hog games, GTX 760 would be ideal... barato ra ni compared sa previous gen... cheaper than the GTX 660 Ti but performance wise, it's on the turf of the GTX 670... mahal pa ang GTX 670 ron pud...

- - - Updated - - -

Cause Fermi has different implementation in terms of cuda vs kepler. Fermi cuda cores =/ Kepler cuda cores. That's apples to oranges comparison ang gibuhat sa reviewer.

the reviewer is just comparing the performance of both architectures from Nvidia... naa man gyud na choices na sila, ang buyer na pili asa iyaha... if naa sya'y Fermi based GPU, usa na's basihan nya to upgrade to a Kepler one... reasonable ba or dili...

for Apple to Apple comparison, atoang basinhan dha ang the same architecture... then ang performance to price ratio ang atoang tan-awon if worth ba to buy let's say... a GTX 680 or its lower Kepler siblings for Adobe Suite...

[werty]

si TS ray makatubag ana for the final specs.. i think he said already about 2 months pa sya mupalit.. by then i think mausab na iya budget, bsin 100k na budget

[lloyd_joy]

^ hehehehe... mao gyud... by the way, sus, kun mo support lang unta ang Premier ug After Effects ug AMD Stream, kani, mas nindot gyud unta kay mas power ni sya in crunching data compared sa Nvidia Kepler (no idea ko's data crunching performance GK110 sa Kepler... atoa i-research beh)... pero, wla man gyud... Photoshop ra ang ning support sa Stream...

mura ang upcoming version sa Adobe for video kay mo support na'g AMD Stream noh?

[werty]

no hands-on experience pud ko ana na field. suggest ra ko on what others have tried hehe...

[Sempron Guy]

For example, the GTX 570 video card has 480 CUDA cores, while the GTX680 has 1,536 CUDA cores. Since the GTX680 has more than 3X the number of CUDA cores over the GTX570, you might expect the GTX680 to be 3X faster than the GTX570. But, they are not. They are only few seconds difference. The reason is, the memory interface width is smaller on the GTX6XX series of video cards, even though they have more cuda cores. Where the GTX5XX series of video cards have a much wider memory interface width, which makes up for them having fewer cuda cores.

I'm referring to this in particular since mao ni ang murag base sa iyang article. I think ang sakto siguro nga wording kay, "More doesn't always means better if we are comparing different gen. cards." That is for the current gen vs past gen comparison. For apples-apples, sakto ang analogy since adobe is purely cuda based and gamay ra kaayo ug difference in terms of cuda count ang GTX 660 and GTX 680.