I'm thinking of getting i7 quad core sandy bridge or i5 quad core sandy bridge. Not sure how passmark works and how to read it but saying i5 quad core 5xxx and i7 quad core is 9xxx. And the highest 10xxx



I was thinking of getting this http://ark.intel.com/products/52208/...up-to-3_30-GHz) or this http://ark.intel.com/products/52213/...up-to-3_80-GHz)

But not sure how much faster ? Probably not much may be 10% faster but not much.

Or if both are not that fast I may pay alot and get Intel Core i7 980X .

I'm going to keep this simple...

Most linux programs render video with CPU.

Best to keep in mind what video card rendering video is for such as Ati or Nvidia can play a part in the rendering of video beyond just the CPU time...

Since Linux doesn't have quality drivers for ATI ,best bet is Nvidia rendering...

Of course there are other issues for video rendering to consider already mentioned by TobiSGD...

Still my bigger concern with video rendering is Bus on the motherboard how much data it can handle at one time for the process... Usually to handle a different CPU there is already a given standard.. But usually High performance Motherboard makers have setups to handle cutting edge bus architecture. Within 2 years the high performance board is now at best average.


So best to shop around find the best deal.. Do away with the WHAT IFS... I would also consider a GDDR5 video card for lower GPU power consumption...

:| The highest score is not 10XXX. Its 14XXX from a i7-3960X.

I'll say it one more time- passmark is an artifical benchmark. Passmark scores can be misleading. Look for real world benchamrking using whatever video editing program you are using.

*blinks* Did you even read post #10? The 980X benchmarks well, but can be slower than a i7-2600, or even a i5-2500.

Its been superseded by the i7-3930K, which costs about the same and should be a little faster anyway.

Getting a video card will _only_ help if the program will use the GPU. If it wont, getting a fast video card will just cost money, suck power and produce heat with no improvement.

GDDR5 is lower power consumtion that GDDR3, but video memory is not a major user of power as far as video cards go. You are still dealing in 'what ifs' (does the program that nec207 is using even use the GPU?). Without knowing for sure if the GPU will make any differnce, suggesting a GDDR5 video card is just going to complicate matters.

Umm...ATI/AMD divers, both open and closed source, havent given me any issues with linux.

ASUS ENGT440/DI/1GD5 GeForce GT 440 (Fermi) 1GB 128-bit GDDR5 PCI Express 2.0 x16 HDCP Ready Video Card

Video cards draw power or u wouldn't see 600 watt and up power supplies.. Read comments to this card u will find Person Using a 300 watt power supply to run the card. That is what I look for is user comments not some tech company.

Also I'm running a EVGA 512-P3-1241-RX GeForce GT 240 512MB 128-bit GDDR5 Video Card for 3 yrs No noise or problems with games running in high settings in Windows 7 also video editing in Arch linux, Linux Mint...
I don't favor ATI over Nvidia. I just know Nvidia cards have been working in Linux for a long time. I have seen in the past 2 yrs 100% increase in compatible ATI cards...

Look I have seen ATI vs Nvidia for just about as long as AMD vs Intel.. Give it six months and we will have Amd cpu out preform the Intel chip.

Furthermore what I was getting at was nec207 never specified just animated video editing or camcorder video editing. There are differences in taking advantage of Shading and pixel editing 2D/3D animation...

Last what ultimately I was getting at is both Nvidia and ATI have huge video card sales this time of the year. That could drop the price down to 50 to 60 dollars for GDDR5 card. Making it a great steal for such a card no matter the manufacturer. I had already seen ATI cards running 50 bucks with instant and manufacturer rebate on Newegg.com less then a month ago.

IMO I go with what is mid level Video Cards and Upper level CPU weather it is ATI/Nvidia/AMD/INTEL... Do the research and find the best bang for the buck

Me I bought a Compaq with AMD Athlon(tm) II 170u Processor 2ghz 20 watt power consumption for $250.. Put my Nvidia card in the system. Left me with enough money to buy a 32" LED TV by VIZIO 1080p got the model at 289 when it retails at 400.. Also bought a wireless keyboard and mouse for 30 bucks. So I think I'm sitting pretty sweet on deals... In my futon chair.(futon chair was outrageous price. Cant bare the thought)

This is not my work horse for video editing...

If you need power calculations, I would use:
http://www.extreme.outervision.com/p...ulatorlite.jsp
and then add 50 to 100 watts just to be sure and plan ahead.

Of course they draw power, but the main cause for that is the video chip, not the RAM. Yes, GDDR5 needs less power, but that will be pretty unnoticeable if you have a highend-videochip.

With a GT240 you will not be able to play recent AAA games in high settings.

Definitely not. AMD has currently no architecture available (including the upcoming Piledriver) that can outperform Intel's Ivy Bridge architecture (or even the older Sandy Bridge), and they will not magically come up with a faster architecture in the next six months.

Yes, there are differences. But they totally depend on the used software. Before we know which type of video editing the OP is doing and which software he uses there is no point in recommending a specific videocard.

Benchmarks divide into two-and-a-half fundamental types

• real world benchmarks measure the performance of some kind of real world task
• synthetic benchmarks measure performance doing something that no one really wants to do in this 'real world' thingy; you know, stuff like 'bandwidth to memory', 'data rate from the hard disk', or something
• 'somewhere in-between' benchmarks that do some kind of hybrid of the previous two

The 'real world' ones could be useful if you have a real world benchmark measuring the actual task that you want to carry out. Otherwise, it could be slightly informative or completely deceptive. And, I've even seen things like benchmarks on similar sounding tasks on two different databases, where the results from one show little correlation with the results on the other. So, while you might think, in this context, one database is pretty much like another, and they are both doing similar operations, they are both using SQL, so the results ought to be similar, it ain't necessarily so.

Synthetic benchmarks could be useful, if, for example, you have a slow system and wanted to know if making some particular change could be constructive. Even then, you have to have intelligence, knowledge and some feel for how things work. You may think that this means that they are not for you, in this present case, but most people want the information for something stupid like 'bragging rights' or what is charmingly known in this industry as 'willy waving', but it could indicate to you whether extra memory bandwidth has the possibility of improving matters (or would you be better with lower latency?).

Passmark has currently no use to me, and possibly none to you. That doesn't mean that it is completely without use, but, if their website doesn't give more information then it is going to be difficult.

Moore's law is not a law of physics. These days it has become an indication - a rule of thumb - to semi manufacturers of how much they need to spend on new fabs in order to stay on the pace, because it is an indication of how much progress their competition will plan on having.

Even if this version of Moore's law was accurate, which it isn't, your maths is surprisingly random. Three years = 2 x 18 months, and 2 x 2 is four and not three. 2010 - 2005 is five years, or sixty months. Sixty divided by eighteen is 3.3, and even if we call that 3, for easy mental arithmetic, that would be 2 x 2 x 2 (or eight, for moderate values of 2) times as fast and not 'no more than 3'.

Now, a certain amount of this thread seems to be predicated on the assumption that there is a number that describes how much faster one cpu is than another. There isn't. No such single number. There will be a range, and that range might be something a little difficult like 0.8 - 1.4. So, that would mean that processor two could be quite a bit faster or a bit slower than processor one, which does make it a little difficult to know where to place your money. You wouldn't want to buy the higher-priced one and find out that it was slower in your use case, but that could happen.

OTOH, if you had lots of different benchmark results, and they ranged from 1.25 - 1.65 and grouped around 1.45, you might not know exactly how much faster it would be for you, but it seems you could expect at least 1.25 (with a few caveats) and more likely something like 1.45.

This may or may not make it worth spending the money, but that's really not my problem.

Oh, and do bear in mind that having a faster cpu doesn't make your network connection faster, your hard (noticeably) disk faster, your ram larger, so if your task is limited in speed by one of these other factors, spending money on your cpu isn't going to do much for you, unless you also deal with the actual bottleneck.

Quote:

*blinks* Did you even read post #10? The 980X benchmarks well, but can be slower than a i7-2600, or even a i5-2500. Its been superseded by the i7-3930K, which costs about the same and should be a little faster anyway.

Yes but if I understand you are saying you cannot go by the passmarks do to it is not 100% real !! Meaning for passmarks to be real you need to look for passmarks on video editing program I will use and CPU I may want, and same RAM ,MOB and bus speed ,hard drive so on .

You saying with these fake passmarks you cannot come to the conclusion that one has 1,000 points and other 5 ,000 points that the CPU with 5 ,000 is 5 times faster unless they use real passmarks same video editing software , same RAM ,MOB and bus speed ,hard drive so on.

All the passmarks do is tell you the CPU is faster but with out real passmarks you will not know how much faster be it 10% , 20% , 50% or 3 times faster.

Now you got it.

Knowing what software you are using may help provide a solution.


I don't think passmark is fake as such. It is a test like any product might be tested. You use it as an indication of performance. Like saying price is an indication of performance. Not always exact but a guide.

Most of those tests are done in a windows world too. The results may be very different in linux. As above it is noted that the software generally uses cpu in common linux tools where the more modern systems in windows use GPU.

Closer, but I'm going to try an analogy (I hate this, but sometimes...):

Imagine you want a car. Imagine that you want a (moderately) fast car. You see a car with a 0-60 time of eight seconds and another with a 0-60 time of six seconds. Obviously, the one with the six second time is the faster car, yes? Well, it depends what you mean. If you want the one that goes around corners fastest, the 0 - 60 time tells you nothing about that. If you want the one that has the highest top speed, the 0 - 60 time really doesn't tell you that either. If you just want to tell your brother-in-law that you have a faster car than his, then anything might do

It not uncommon for video editing for one hour HD video to take 4 or 5 hours !! Now if i5 2400S does it in 4 hours and the i7 2600 does it in 3 hours and 30 minutes the i7 2600 is not that much faster.

If the i7 2600 does it in 3 hours or 2 hours and 30 minutes than yes the i7 2600 is much faster.If the i7 2600 does it in 2 hours or 1 hour and 30 minutes than it is really fast.

But from what I read the i7 2600 may be not that much faster than the i5 2400S .So this thing I do not want spend the money if it is only be bit faster .

If i7 2600 does it in 3 hours or 2 hours and 30 minutes than yes I may get it !!! now if the i7 2600 does it in 2 hours or 1 hour and 30 minutes for sure I will get it !!

But if the i7 2600 is not that fast and does it in 3 hours and 30 minutes the i7 2600 is not that much faster.

I think the i5 is the best value, that's what I would buy. I've also read that if you replace the very small standard fan and thermal paste with a better fan, you can easily overclock it and get the same performance as an i7. Not that I would really overclock it, but I would replace the fan and paste.

1 members found this post helpful.

What video editing software are you going to use?

+1 for that.