New PC Build

[faramith]

Im thinking in making a new PC with the following:


Motherboard ASUS P6T Deluxe V2 intel x58
cpu i7 920 or 950
ram mushkin 6GB (3 x 2GB) 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800)
graphics card SAPPHIRE 4870 2gb gddr5 pci-e 2.0

operating system INTEL SSDSA2MH160G2C1 X25M 160GB SATA 3.0 Gen 2, MLC, 2.5" SSD
2 velociraptor 300gb western digital on raid 0
2 1TB western digital on raid 1 for file backups

I read that SATA3 its just around the corner, that would improve performance
when editing video right? if so should I wait a bit for it to be implemented in Motherboards
or no?

Also This little guy here:
http://www.ocztechnology.com/products/solid_state_drives/ocz_z_drive_m84_pci_express_ssd
promises to be an excelent candidate to replace the velociraptor 300gb raid 0 I was thinking in getting,
since its a PCI-e SSD I find it amazing, also its going to cost about $699 for the 256gb one..

any suggestions are welcome! thanks

[jagabo]

I read that SATA3 its just around the corner, that would improve performance when editing video right?

No. Drives aren't even anywhere near using the full bandwidth of SATA 2 yet. And unless you're working with uncompressed video drive speed is pretty much irrelevant when editing and encoding video.

Copying big files from one drive to another is another story. Faster (greater sustained throughput) drives will make a difference there. But the rate the drives can get data on and off the platters is much lower than the SATA 2 interface. In fact, they're just now getting to the point where they need more than a SATA 1 interface.

[redwudz]

Two problems with SSD drives:

Turn off defragging for the drives, it will damage them as it repeatedly uses the same cells and they have a limited lifespan. Vista does auto defragging, so if you use that OS, you need to disable it for the SSDs. I assume Windows 7 does the same.

They function best at about 50% free space as the drive needs room to move the files around in lieu of defragging, so this limits their usable space.

[rallynavvie]

Usage for this machine would help.

Despite all the people who say RAID 0 is great know that you are effectively doubling the chance of that volume failing. If one drive goes you lose the data on both drives. It's great for scratch disks or testing things but not on any volume that important. If you don't care about losing that data and still want the performance increase then go ahead. Run the 300GB drives separately and you'll probably get more use from them.

As to SSD I won't touch the things until they all support TRIM. That increases the life of the SSD (remember that SSD have almost half the MTBF of a consumer rotational drive) and greatly increases its performance. And it certainly isn't worth the price-to-performance right now for a consumer. And why are they marketing that PCI-E SSD as "taking the SATA bottleneck out of the equation with a high-speed PCI-Express interface"? You realize your SATA controller uses the same ort of serial link to your chipset, right?

[Video Head]

Editing is not the time consuming part, encoding is. I make my systems work at night while I am cozy and comfy or watching a game. Cruel, I know, but a good use of my time.

$699 for 256GB? You can buy alot of SATA disc for that money - even build an entire NAS box. But go ahead, we need early adopters to buy it so that the price can come down.

[faramith]

After reading bad things about SSD I think im better off with normal sata drives, I want an encoding/editing pc. I need a setup for editing big files on a fast HDD or fast raid setup any pointers? after the file is done I'll just move it to a Raid 1 backup drive where I can finish up the project. thanks

[rallynavvie]

Yes, run those two 300GB Raptors separately. You don't want to be doing editing work on a drive that your OS is running on, even if you are running them in stripes. Use one of the Raptors for Windows and applications and use the other as your project drive. Once done editing transcode down to the end format to your larger storage drive then transfer whatever project files you need to keep over to that drive to free up space for the next project.

As was said above you don't need anything more than a regular SATA drive for consumer video no matter how large. I was working with 10GB DV AVI files on IDE drives several years ago with no issues. Even HD formats now don't need anything more than SATA. It's when you get to uncompressed formats (like Redcode) that you need more power from your storage.

[jagabo]

Big, inexpensive SATA drives have greater sustained read/write speeds than the small Raptors.

http://www.pcstats.com/articleview.cfm?articleid=1746&page=5
http://www.cdfreaks.com/review/15169-Samsung-Spinpoint-F1-HD103UJ-review/Reading-and-W...Performance-2/

Seek time on the Raptors remains much faster though.

[rallynavvie]

Big, inexpensive SATA drives have greater sustained read/write speeds than the small Raptors.

http://www.pcstats.com/articleview.cfm?articleid=1746&page=5
http://www.cdfreaks.com/review/15169-Samsung-Spinpoint-F1-HD103UJ-review/Reading-and-W...Performance-2/

Seek time on the Raptors remains much faster though.

Yeah the high-density platters of the larger drives means more data is going past the head, but that's sustained I/O which is most useful for backups and linear I/O. Encoding from one drive to another would utilize this.

However for actual editing the faster seek times are going to be much more useful, especially as you use several resources for your project.

[ocgw]

Im thinking in making a new PC with the following:


Motherboard ASUS P6T Deluxe V2 intel x58
cpu i7 920 or 950
ram mushkin 6GB (3 x 2GB) 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800)
graphics card SAPPHIRE 4870 2gb gddr5 pci-e 2.0

operating system INTEL SSDSA2MH160G2C1 X25M 160GB SATA 3.0 Gen 2, MLC, 2.5" SSD
2 velociraptor 300gb western digital on raid 0
2 1TB western digital on raid 1 for file backups

I read that SATA3 its just around the corner, that would improve performance
when editing video right? if so should I wait a bit for it to be implemented in Motherboards
or no?

Also This little guy here:
http://www.ocztechnology.com/products/solid_state_drives/ocz_z_drive_m84_pci_express_ssd
promises to be an excelent candidate to replace the velociraptor 300gb raid 0 I was thinking in getting,
since its a PCI-e SSD I find it amazing, also its going to cost about $699 for the 256gb one..

any suggestions are welcome! thanks

Take a look @ the Samsung Spinpoint F3 1TB HDD, $80USD each and they are faster @ sustained read and writes than the WD Velociraptor

http://www.bit-tech.net/hardware/storage/2009/10/06/samsung-spinpoint-f3-1tb-review/3

The F3 has (2) 500GB platters, the 500GB version of the F3 is a single platter HDD for $55USD

(3) Spinpoint F3's in a Raid5 config would be as fast as a pair of Velociraptors in RAID0, offer fault tolerance and redundancy, and have 5 times the volume for less money and would make less noise

I just bought a single F3 as my new boot drive for my 10TB HTPC/Home Media Server and am very pleased w/ it

I had 1st gen 10K Raptors in Raid0, Fujitsu 15K SAS in RAID0, but when the large platter, high density SATA II HDD's emerged they more or less made the small capacity high speed spindle enterprise class HDD solution for high performance desktops obsolete

in my most humble of opinions

ps. 5 of 5 eggs @ newegg too lol

ocgw

peace

[rallynavvie]

No, they still do not offer fast seek times which you need for random I/O. There are plenty of tasks where faster rotational speed drives are going to perform better. As I said above sustained I/O transfer speeds are great for linear read/write like backups and copying between drives. Faster RPM drives were never meant to provide faster sustained I/O, in fact there are 5400rpm drives that have better sustained I/O speeds than some 10krpm drives. It has always been meant to decrease seek times for random I/O (DB and some DCC work). It was purely marketing and fallacy that perpetuated the myth that higher-rpm drives were overall the better choice.

They are two different tools for two different things. Faster seek times for better random I/O such as your operating system, applications, and databases. Faster sustained I/O for linear work such as copying, transcoding, and backups.

I'd like to add that while you can use certain RAID levels to speed up your sustained throughput speeds you cannot increase the performance of your drives' seek times.

[ocgw]

No, they still do not offer fast seek times which you need for random I/O. There are plenty of tasks where faster rotational speed drives are going to perform better. As I said above sustained I/O transfer speeds are great for linear read/write like backups and copying between drives. Faster RPM drives were never meant to provide faster sustained I/O, in fact there are 5400rpm drives that have better sustained I/O speeds than some 10krpm drives. It has always been meant to decrease seek times for random I/O (DB and some DCC work). It was purely marketing and fallacy that perpetuated the myth that higher-rpm drives were overall the better choice.

They are two different tools for two different things. Faster seek times for better random I/O such as your operating system, applications, and databases. Faster sustained I/O for linear work such as copying, transcoding, and backups.

I'd like to add that while you can use certain RAID levels to speed up your sustained throughput speeds you cannot increase the performance of your drives' seek times.

No?

rally my friend, I clearly stated the F3 is faster @ sustained read and writes, never claimed they had faster seek times

http://www.bit-tech.net/hardware/storage/2009/10/06/samsung-spinpoint-f3-1tb-review/3

all depends on what you need more, I am shoving around blu rays @ 20-30GB a pop, sustained writes and reads is what I can use the most

btw marketing myths not withstanding I used to get into BF2 rounds so fast (when the maps changed) I could run over to the other side and steal their chit before the other team spawned in when i got my 10K Raptors in 05', the 15K Fujitsus in 07' were so after all the other guys got raptors I could get a plane w/o getting shot in the back lol

Lieutenant General |SI| SERIALBUTCHER

ocgw

peace

[rallynavvie]

when the large platter, high density SATA II HDD's emerged they more or less made the small capacity high speed spindle enterprise class HDD solution for high performance desktops obsolete

That is the statement I had issue with. I had already stated further up about the dense platter vs. high RPM drives with their pros/cons.

[ocgw]

when the large platter, high density SATA II HDD's emerged they more or less made the small capacity high speed spindle enterprise class HDD solution for high performance desktops obsolete

That is the statement I had issue with. I had already stated further up about the dense platter vs. high RPM drives with their pros/cons.

Well I did say for high performance desktops, not servers or workstations, guess it depends on how you define "desktop" I guess?

maybe I should have said home or personal PC, especially gaming (games that load large maps) or HTPC rigs

ps. I do understand your reasoning sir

ocgw

peace

[faramith]

Ok so what HDD would be best for Encoding files,and wich one for handling big files in for example Video Studio where you have to scroll thru the movie, and wich one to transfer from one HDD to another? Im at a loss here.. thanks!

[rallynavvie]

You'll want a drive with higher RPM (the Raptor) for editing and working on your video project and you'll want the big, high-density drives for storage. The bottleneck when encoding will be your CPU, not your HDD speed so don't worry about that as long as you're reading from one drive and writing to another as you encode (otherwise you get a lot of concurrent read/write on the same drive which can slow you down).

The workflow should essentially go like this:
Capture/record source video to your project drive (10krpm SATA)
Edit and master your video on your project drive
When the project is finished do your encodes from that project drive to your storage drive (big 7200rpm SATA)
Copy any of the project files you want to back up off from your project drive to your storage drive
Clean up any old project files to prepare for your next video project then start over

[ocgw]

You'll want a drive with higher RPM (the Raptor) for editing and working on your video project and you'll want the big, high-density drives for storage. The bottleneck when encoding will be your CPU, not your HDD speed so don't worry about that as long as you're reading from one drive and writing to another as you encode (otherwise you get a lot of concurrent read/write on the same drive which can slow you down).

The workflow should essentially go like this:
Capture/record source video to your project drive (10krpm SATA)
Edit and master your video on your project drive
When the project is finished do your encodes from that project drive to your storage drive (big 7200rpm SATA)
Copy any of the project files you want to back up off from your project drive to your storage drive
Clean up any old project files to prepare for your next video project then start over

I disagree, if you edit and master your project on one drive it will be do simutaneous read writes to itself, depending on what platters that it access that can be a big gain or a bottle neck

If you read and write to same platter speed is greatly reduced, read from one platter and write to another on the same HDD = super fast transfer

I play it safe and always read from one, and write to another high sustained average read and write speed HDD's for all editing and encoding operations, not dealing w/ many small fileS so "to heck" w/ "fast seek times"

Fast seek times rule, if you are a corporate data server LOL

Still nice to have the OS/apps on a nice Raptor or SAS RAID array, or an SSD tho', I was a huge advocate of enterprise drives in high performance desktops, but I gave it up for max storage myself

I own (2) 10K WD Raptors, and (2) 15K Fujitsu SAS HDD's which I ran in RAID0 for OS and apps, loved them, but when high density drives caught up to the small enterprise drives in transfer speed I started using them for storage, now that they are leaving even Velociraptors "in the dust" w/ sustained reads and writes the smaller enterprise drives are "paper weights" IMO, and the high density SAS drives are prohibitively expensive

I suggest using all large platter high sustained average transfer drives, and always write from one drive to another drive every operation, I go back and forth, and back and forth again between my drives so I am rarely ever reading and writing on the same drive

ocgw

peace

[rallynavvie]

I disagree, if you edit and master your project on one drive it will be do simutaneous read writes to itself, depending on what platters that it access that can be a big gain or a bottle neck

If you read and write to same platter speed is greatly reduced, read from one platter and write to another on the same HDD = super fast transfer

I play it safe and always read from one, and write to another high sustained average read and write speed HDD's for all editing and encoding operations, not dealing w/ many small fileS so "to heck" w/ "fast seek times"

Don't confuse encoding with editing, they are two entirely different things. You edit source video until you have your final project then you encode that project into your single final video file. The source file(s) are generally going to be large, uncompressed video which you'll need faster seek/access times due to jumping around within that file, especially if you have one giant source file you're cutting/splicing apart. When editing you aren't really doing linear I/O because of the nature of the work. Combine that with multiple large video files and the issue is compounded. Fast seek times isn't all about small files all over your drive it's also about moving between places on very large files. A giant source video file can be looked at as a giant database file of image data. And there is very minimal writing when doing the editing because the project file stores in/out points, effects, cuts, and all your edits as part of its data; it isn't authoring the file as you're editing. That would be inefficient. It stores that information until you're ready to master the file which is done as part of the encoding process. The actual writing done to that file is pretty minimal.

If you'll notice my workflow I do say to encode from one drive to another and I guarantee your CPUs are going to be the bottleneck when encoding so you don't need assloads of bandwidth on your drives to support what the CPU can crunch. Not even my rig can top out any of my drives' I/O while encoding video. I also mentioned the dangers of encoding to/from the same drive which is why I always recommend a second destination drive when encoding.

Now if people are considering "editing" to just be cutting the intro/outro titles from a ripped DVD/BRD source file prior to encoding to a smaller format then seek times are irrelevant, but I'd hardly consider that video editing.

[ocgw]

I disagree, if you edit and master your project on one drive it will be do simutaneous read writes to itself, depending on what platters that it access that can be a big gain or a bottle neck

If you read and write to same platter speed is greatly reduced, read from one platter and write to another on the same HDD = super fast transfer

I play it safe and always read from one, and write to another high sustained average read and write speed HDD's for all editing and encoding operations, not dealing w/ many small fileS so "to heck" w/ "fast seek times"

Don't confuse encoding with editing, they are two entirely different things. You edit source video until you have your final project then you encode that project into your single final video file. The source file(s) are generally going to be large, uncompressed video which you'll need faster seek/access times due to jumping around within that file, especially if you have one giant source file you're cutting/splicing apart. When editing you aren't really doing linear I/O because of the nature of the work. Combine that with multiple large video files and the issue is compounded. Fast seek times isn't all about small files all over your drive it's also about moving between places on very large files. A giant source video file can be looked at as a giant database file of image data. And there is very minimal writing when doing the editing because the project file stores in/out points, effects, cuts, and all your edits as part of its data; it isn't authoring the file as you're editing. That would be inefficient. It stores that information until you're ready to master the file which is done as part of the encoding process. The actual writing done to that file is pretty minimal.

If you'll notice my workflow I do say to encode from one drive to another and I guarantee your CPUs are going to be the bottleneck when encoding so you don't need assloads of bandwidth on your drives to support what the CPU can crunch. Not even my rig can top out any of my drives' I/O while encoding video. I also mentioned the dangers of encoding to/from the same drive which is why I always recommend a second destination drive when encoding.

Now if people are considering "editing" to just be cutting the intro/outro titles from a ripped DVD/BRD source file prior to encoding to a smaller format then seek times are irrelevant, but I'd hardly consider that video editing.

I am certainly not confused, been makin' home movies (and a few weddings) w/ non-linear editing for 15 years, I am saying you over emphasize the advantages of seek time vs sustained transfer rate, let's put it in simple terms, 5ms vs 11 ms seek time would be an advantage in video editing over sustained average transfer rate if you are splicing 1 frame @ a fvckin' time

Let me put it another way, as you set edit points, any drive is faster than your editing input, after you start the actual transfer unless you are editing 1 frame @ a damn time transfer rate rules

That's just w/ editing, even when it comes to encoding the cpu will be the bottleneck w/ some encoding (higher quality HD) for instance, but not necessarily for SD sources, that's why my rig topped (and every one elses rig) your rig in fps on your own freakin' encoding test on the 1st pass

*superior i/0 performance*

Seek time is a huge advantage when 100 ppl are accessing your database, can't really have a huge impact on a single user enviroment for video, gaming maybe

ocgw

peace

[rallynavvie]

as you edit, any drive is faster than your editing input, after you start the actual transfer unless you are editing 1 frame @ a damn time transfer rate rules

What "actual transfer" are you referring to? Which editing process are you referring to there?

Where did I benchmark my machine on video encoding?

[ocgw]

as you edit, any drive is faster than your editing input, after you start the actual transfer unless you are editing 1 frame @ a damn time transfer rate rules

What "actual transfer" are you referring to? Which editing process are you referring to there?

Where did I benchmark my machine on video encoding?

After you set your edit points (which I contend any modern drive can write faster than you can input) and you start the process if you set your output to a different drive it is a transfer

OMG you don't remember everyone here posting their times in your benchmarks test and I had the highest fps (only one w/ 80+ fps) on the first pass beating everyone here including you and everyone else w/ dual quad Xeons?????

Superior i/o will speed up editing, even encoding if it isn't too cpu intensive (I certainly didn't get top 1st pass times because of my AMD quad performance)

Edit: Sorry, the Techarp test that Graysky posted

https://forum.videohelp.com/topic347859.html

Not only am I not "confused", I got damn good memory

ocgw

peace

[lordsmurf]

been makin' home movies (and a few weddings) w/ non-linear editing for 15 years,

You've been editing video since 1994? I find that hard to believe.

[ocgw]

been makin' home movies (and a few weddings) w/ non-linear editing for 15 years,

You've been editing video since 1994? I find that hard to believe.

Before PC's they had editing VCR's for non-linear editing, I didn't have full blown multi-monitor/multi-source analog video editing station, but I did non-linear video editing w/ my Hitachi Hi-8 camcorder and my Mitsubishi S-VHS editing VCR

This is news!?

I paid $800USD for my $1200USD MSRP Mitsubishi Super-VHS editing VCR, but it might have been 95' or 96', I would have to find my receipt if I still have it

In the early 90's us "old timers" had some pretty nice gear, Hi-8 camcorders, Laser Disc Players, Pre-Pros w/ Dolby Digital, and medium format cameras

I started digital editing w/ my purchase of a firewire card for my old 800Mhz Dell PC

ocgw

peace

[faramith]

How can you read from a platter and write to another platter in the same HDD? Im getting confused with all this talk @_@, so far .. what I gather is getting a Raptor 300Gb for OS. Raptor 300GB for editing and when I have the editing finished then I encode to another HDD? will it encode faster than encoding to the same Raptor? since my project HDD will be a Raid 1 1TB WD HDD so I dont loose the project, I think encoding to this array will slow down the process no?.

[jagabo]

How can you read from a platter and write to another platter in the same HDD?

You have no control over this. Ignore it.

what I gather is getting a Raptor 300Gb for OS.

That will save you ten seconds at boot time. Some applications will load faster.

encode to another HDD? will it encode faster than encoding to the same Raptor?

Yes, but by how much is negligible unless you are working with uncompressed or losslessy compressed video. Going from DV to MPEG for example it will only save you about a minute on an hour long encode. DV to DV will be more noticeable though not huge, in the 10 percent range. Uncompressed RGB to uncompressed RGB will be be large.

since my project HDD will be a Raid 1 1TB WD HDD so I dont loose the project, I think encoding to this array will slow down the process no?.

No.

[ocgw]

How can you read from a platter and write to another platter in the same HDD?

You have no control over this. Ignore it.

what I gather is getting a Raptor 300Gb for OS.

That will save you ten seconds at boot time. Some applications will load faster.

encode to another HDD? will it encode faster than encoding to the same Raptor?

Yes, but by how much is negligible unless you are working with uncompressed or losslessy compressed video. Going from DV to MPEG for example it will only save you about a minute on an hour long encode. DV to DV will be more noticeable though not huge, in the 10 percent range. Uncompressed RGB to uncompressed RGB will be be large.

since my project HDD will be a Raid 1 1TB WD HDD so I dont loose the project, I think encoding to this array will slow down the process no?.

No.

You read from 1 platter and write to another on the same drive when you have 2 or more partitions on the same drive and you write from 1 partition to another

You do have control over it, you prevent reading and writing to the same drive when editing video by having different source and destination drives

And you can use it to your advantage if you set up your partitions right

Even if you use 1 project drive use seperate partitions so you are writing from 1 part of the drive to another

On a unfragmented drive you are not necessarily gonna' boot faster w/ a Raptor than w/ a faster sustained average read speed high density drive

The day of the Raptor is over, 5 years ago they had top sustained transfer speeds and great seek times, now that they have fallen behind in transfer rates and are losing the $ pr gigabyte war they would be finished for high performance desktops if it weren't for the myth of high rpms equaling high performance

Also you got it backwards jagaboo the lower bit rate video which is less cpu intensive makes a faster i/o even more of an advantage, uncompressed video would make the cpu the bottleneck making no difference how you do it

To farasmith: listen to who you want I don't care, I got straight A's in college electrical, electronic and computer classes, I know what the "F" I am talkin' about, not guessing

ocgw

peace

[rallynavvie]

After you set your edit points (which I contend any modern drive can write faster than you can input) and you start the process if you set your output to a different drive it is a transfer

OMG you don't remember everyone here posting their times in your benchmarks test and I had the highest fps (only one w/ 80+ fps) on the first pass beating everyone here including you and everyone else w/ dual quad Xeons?????

Superior i/o will speed up editing, even encoding if it isn't too cpu intensive (I certainly didn't get top 1st pass times because of my AMD quad performance)

Edit: Sorry, the Techarp test that Graysky posted

https://forum.videohelp.com/topic347859.html

Not only am I not "confused", I got damn good memory

The "output" process you're referring to is the transcode/encode which I already stated gets better performance drive-to-drive. Actually playing with splicing and edit points and then doing a rough playback will jump around as much as you splice it. If you have minimal amounts of cuts this isn't a big deal but when you're taking a 100GB HD AVI file and dropping edits all over the place (like if you're a cut-happy director like Aronofsky) your read heads go nuts jumping around the source. But all this arguing is moot since it seems like neither of us is going to yield. I take issue with the "I know more than you" stuff though because I did run a successful DVD mastering business for years back in the days when you needed Maestro (stupid dongle and all) to properly author a DVD and nobody was yet trained on how to put them together. I don't just go dropping that information all the time to back up what I'm saying though. I would hope my elocution and radical conjecture is enough

FWIW I did not participate in grayskys x264 benchmark because it didn't run all 8 cores when it encoded. So yes, you probably did see some Harpertown machines on there running similar benches to any other desktop quads and I would expect the higher-clocked desktop quads do well. Especially if they've been overclocked. I don't see why this is so surprising?

[ocgw]

After you set your edit points (which I contend any modern drive can write faster than you can input) and you start the process if you set your output to a different drive it is a transfer

OMG you don't remember everyone here posting their times in your benchmarks test and I had the highest fps (only one w/ 80+ fps) on the first pass beating everyone here including you and everyone else w/ dual quad Xeons?????

Superior i/o will speed up editing, even encoding if it isn't too cpu intensive (I certainly didn't get top 1st pass times because of my AMD quad performance)

Edit: Sorry, the Techarp test that Graysky posted

https://forum.videohelp.com/topic347859.html

Not only am I not "confused", I got damn good memory

The "output" process you're referring to is the transcode/encode which I already stated gets better performance drive-to-drive. Actually playing with splicing and edit points and then doing a rough playback will jump around as much as you splice it. If you have minimal amounts of cuts this isn't a big deal but when you're taking a 100GB HD AVI file and dropping edits all over the place (like if you're a cut-happy director like Aronofsky) your read heads go nuts jumping around the source. But all this arguing is moot since it seems like neither of us is going to yield. I take issue with the "I know more than you" stuff though because I did run a successful DVD mastering business for years back in the days when you needed Maestro (stupid dongle and all) to properly author a DVD and nobody was yet trained on how to put them together. I don't just go dropping that information all the time to back up what I'm saying though. I would hope my elocution and radical conjecture is enough

FWIW I did not participate in grayskys x264 benchmark because it didn't run all 8 cores when it encoded. So yes, you probably did see some Harpertown machines on there running similar benches to any other desktop quads and I would expect the higher-clocked desktop quads do well. Especially if they've been overclocked. I don't see why this is so surprising?

1. You don't have to encode or transcode when you edit, that is an assumption on your part, you can do each of those tasks seperately going from 1 drive to another and back, just edit drive to drive, and after that if you need to transcode do that in a seperate step drive to drive, better yet if you are editing from 2 cameras have 2 seperate input drives and a final output drive same as w/ a good analog video editing station, this ain't new, then encode or transcode into your desired filetype

2. Graysky's benchmark does utilize all 8 cores, that's why the dual quad Xeons "dusted" my single quad in the cpu intensive 2nd pass, I only beat them on the first pass because it wasn't cpu limiting and I had superior i/o performance

Here again you just don't see where the bottleneck is, you are a very knowlegable guy, but I am sorry, you are not the hardware expert here

How exactly does "running" a company make you an expert? You should have hired a hardware expert

For you information your heads are banging around because you are reading and writing to the same drive, and probably on the same partition w/ your files all fragmented together

it's called "disc thrashing"

Now you are trying to convince a guy that doesn't know any better to make the same mistakes you made

go ahead faramith, follow rallynavvie's advice, he used to "run" a company lol

I set my edit points on the OS drive, and then splice the video together from a 2nd source drive to a 3rd destination drive, then encode to a 4th drive, then store the video on other drives, I know it is a lot of drives but it isn't expensive if you don't waste all your money on just 2 drives just to say you have 10K Raptors, I put my money where my mouth is, I have ( 8 ) high density HDD's in my main PC ranging from WD 640GB 16mb cache HDD's w/ 320GB platters to Seagate Barracuda 1500GB 32mb cache HDD's w/ 375GB platters

OS/apps on the new Samsung Spinpoint F3 1TB HDD w/ 500GB platters that "blows the doors" off of Velociraptors

https://forum.videohelp.com/topic368691.html

http://www.bit-tech.net/hardware/storage/2009/10/06/samsung-spinpoint-f3-1tb-review/3

ocgw

peace

[faramith]

nice drive the F3, too bad its out of stock on newegg

[ocgw]

nice drive the F3, too bad its out of stock on newegg

They will be back in stock soon, don't worry, they are probably "selling like hotcakes" lol

Funny how that drive is "standing tall" in test after test w/ drives that cost multiples of the F3, that is why I said earlier that you could run a F3 Raid5 array for the cost of a single Raptor

For the cost of 2 Raptors You could have a F3 RAID5 array and a F3 Raid0 array to run processes back and forth between them and know that i/o is never the bottleneck on any operation

ocgw

peace