However with a quick search I am seeing host based and software RAID grouped into tha same cat, but to me its something different
I'm sure, but the reality is that Intel 'host based' RAID does not do anything in the hardware. If it does, they haven't bother letting anyone use it on a non-Windows platform, which is both stupid and not like Intel.
For example, a RAID5 volume on Intel ICH will get ya oh somewhere in tha neighborhood of 20-30MB/s writes on ICH7 or greater
Nice low numbers, there.. I can pour >100MB/s into my software RAID-5 on this machine, and it's bottlenecked by old drives.
However with a quick search I am seeing host based and software RAID grouped into tha same cat, but to me its something different
I'm sure, but the reality is that Intel 'host based' RAID does not do anything in the hardware. If it does, they haven't bother letting anyone use it on a non-Windows platform, which is both stupid and not like Intel.
For example, a RAID5 volume on Intel ICH will get ya oh somewhere in tha neighborhood of 20-30MB/s writes on ICH7 or greater
Nice low numbers, there.. I can pour >100MB/s into my software RAID-5 on this machine, and it's bottlenecked by old drives.
Well im not lieing to ya, I tried, figuring dam well I was wasting my time, to make a 4 drive RAID5 set when I bought this Z77 motherboard, I didnt get no 100MB/s writes, oh ya reads are fine, but ICH just isnt powerfull enuff to cal tha parity writes.. If I remmber correctly it was like 30 MB/s or something.. OH well, my RAID5 volumes are all handled by tha Highpoint 4520 I bought, it has a 900MHZ Marvell 88RC9580 chip to handle all tha parity cal, and SAS 6G, 128 devices with expander.. Its a full hardware solution, and I can create and initialize RAID5 volumes right in its bios, and watch them happen there, go play some dos games, reboot and see were tha volumes at, no drivers needed, no os needed..
There I did a quick benchmark for ya, its a 8 x 1TB seagate/wd drive mix, so RAID5 on tha Highpoint 4520, 8 drives, 7TB unformated space, more then 6TB formated..
You wrote 256MB to a controller with a 512MB buffer..
You wrote 256MB to a controller with a 512MB buffer..
Never even thought really, okay so u want me todo twice its buffer size, or what settings? I will rerun it here..
You wrote 256MB to a controller with a 512MB buffer..
Never even thought really, okay so u want me todo twice its buffer size, or what settings? I will rerun it here..
Try a proper sustained write, for starters. 8GB+.
Well, that test only had up to 2GB in tha drop box, so 2GB transfers it was, seemed to drop tha reads, not writes so much.. I also attached tha HDTach test on tha same RAID5 volume..
I still don't think it's doing a sustained write.
I still don't think it's doing a sustained write.
I wouldnt see why not, there is 900mhz chip at tap.. Looks to me like its doing about 50MB/s write and 80MB/s read per drive, and these are older slower sata2 drives, with new drives should be possible to double those numbers, or at least write anyways, I think tha max bandwith available thru tha ports is 2400MB/s, not sure thru tha chip, I read somehere that its a full sata2 chip with sata3 emulation, meaning it will link to sata3 drives, but tha max bandwith over 4 drives is still sata 3Gx4, so u could only use 2 drives per port at full sata 6G speeds, but no mechanical hard drives get over 300MB/s anywayz..
I still don't think it's doing a sustained write.
I wouldnt see why not, there is 900mhz chip at tap..
... Perhaps because it's doing overlapping I/O with varying transfer sizes?
I still don't think it's doing a sustained write.
I wouldnt see why not, there is 900mhz chip at tap..
... Perhaps because it's doing overlapping I/O with varying transfer sizes?
No, its set to 512/64K, on tha RAID5 volume anywayz, unless u mean something with tha test program..
I still don't think it's doing a sustained write.
I wouldnt see why not, there is 900mhz chip at tap..
... Perhaps because it's doing overlapping I/O with varying transfer sizes?
No, its set to 512/64K
I give up.
Nothing wrong with those results anywayz, it works good, I have copied files from one RAID5 volume to another RAID5 volume at like 360MB/s before in windows, never really played that close of attention to tha speeds.. Its also going thru a intel 24port expander board, but only one channel from tha RAID card is hooked up atm..
Oh great, I think I killed tha thread
.. How long Dave until u expect to be able to bulid your beast?
Oh great, I think I killed tha thread .. How long Dave until u expect to be able to bulid your beast?
About to start building it right now...
New machine is setup and working.
52 seconds on the Handbrake test video file I've been using here. That's at least double the speed on my old machine.
Winner!
That's the only test so far, still have to setup everything on the machine.
And that time is a stock install of everything, no tweaking, no overclocking.
Yay
Typical British reserve showing through there. I'll try harder...
<trying>
Double Yay
Twice as fast. Nice.
I await with baited breath the times with cpu and gpu, if gpu is possible with your work flow.
Noise level, power consumption?
Alexander.
Noise level, power consumption?
Very low noise, practically background in my lab. Wouldn't be able to measure it there.
But hae not have the fans speed up yet.
Could be made lower though.
Power is about 50W idle with internal graphics, 0.4W standby.
Some quick tests in Sony Movie Studio rendering.
Albeit while copying my entire hard drive from notebook to new desktop (HDD) in the background!
Reading and writing to my SSD, rendering a 10 minute .MTS clip from the camera to 12Mbps Sony AVC file.
- CPU only about 11min
- GPU acceleration (Intel HD4000 in the i7) about 10min (real time)
- Intel QSV (Quality) about 12min
- Intel QSV (Speed) about 10 min
So that's basically on par with my old notebook. No real major speed improvement in Sony.
Although the hard drive copying could play a part, but the CPU cores aren't being maxed out, and Handbrake is about the same as before the file copy.
Noise level, power consumption?
Very low noise, practically background in my lab. Wouldn't be able to measure it there.
But hae not have the fans speed up yet.
Could be made lower though.
Power is about 50W idle with internal graphics, 0.4W standby.
Wont keep your feet warm at all.
%-B
Some quick tests in Sony Movie Studio rendering.
Albeit while copying my entire hard drive from notebook to new desktop (HDD) in the background!
Reading and writing to my SSD, rendering a 10 minute .MTS clip from the camera to 12Mbps Sony AVC file.
- CPU only about 11min
- GPU acceleration (Intel HD4000 in the i7) about 10min (real time)
- Intel QSV (Quality) about 12min
- Intel QSV (Speed) about 10 min
So that's basically on par with my old notebook. No real major speed improvement in Sony.
Although the hard drive copying could play a part, but the CPU cores aren't being maxed out, and Handbrake is about the same as before the file copy.
Don't suppose you could try MainConcept AVC/AAC @ 10Mbps and 14Mpbs both with and without that fancy video card that Sagan liked so much?
Don't suppose you could try MainConcept AVC/AAC @ 10Mbps and 14Mpbs both with and without that fancy video card that Sagan liked so much?
Yep, will be trying those again for sure.
The video card will certainly get a trial.
About to edit my (long) build video on the new machine. Have now switched over from the notebook, and have a naked machine. Feels like cleaning your workbench...
Damn, the new machines gives me an error a few minutes into rendering using the Sony AVC encoder. Not happy
Trying the Main Concept one (which is a fair bit slower than real time) now on my latest 60min video.
Damn, the new machines gives me an error a few minutes into rendering using the Sony AVC encoder. Not happy
Trying the Main Concept one (which is a fair bit slower than real time) now on my latest 60min video.
Whats tha error? Software issue, or something hardware not seem right?