• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Post your Roadkill HDD Results!

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
ha thats neat. Seagate Momentus 5400- 3901 🙁 🙁

if there is one thing in this TM8200 i need to replace, its the hdd. but 120gb+ 7200rpm 2.5" hdds are expensive. 🙁 and hard to find.
 
Results for primary PC with 2 seperate RAID 1 arrays on Nvidia 590 SLI RAID controller.

All drives are SATA 2 & results are the average of 3 runs on each actual drive because the mirrored drives showed in the program as logical C:, logical D: & physical 0, physical 1 ... I ran the utility on all 4 drives & noted that results for the drive listed as physical were consistantly a couple hundred points higher on both arrays.

Array 1: 2 x 320gb Seagate 7200.10.s = 9179 (logical C: & physical 0)

Array 2: 2 x 500gb WD 5000AAKS = 10809 (logical D: & physical 1)


Not only are the WD's a lot quieter then the Seagates, they seem to be somwhat faster as well which suprised me a little.

Edit: I assumed that the logical/physical choice in the drop-down referred to the 2 drives in each of the RAID arrays on my main PC, but just for the heck of it I went downstairs & ran it on my wifes PC with a single WD Cavier 120gb PATA drive & saw the same thing so I must be mistaken... however the score was still consistantly a bit higher for the "physical" drive on her pc as well (7204 avg of 3 runs on each) ... anyone have an idea as to why?
 
256MB SD card via USB - 7618. Is this supposed to mean something?

Linear Read: 6.66 MB/s
Random Read: 3.3 MB/s
Access Time 1.01 ms
 
Originally posted by: Madwand1
256MB SD card via USB - 7618. Is this supposed to mean something?

Linear Read: 6.66 MB/s
Random Read: 3.3 MB/s
Access Time 1.01 ms
Click to expand...

Doesn't seem like this benchmark is particularly reliable or consistent. Your random read on a USB card is similar to what people are seeing on 7200RPM SATA2 drives. Minerva's 8 drive RAID only hits 9MBps random read.

And with a linear read score like that, how could it possibly get such a high final score? It seems like maybe it's using the total drive size as well to calculate a final score, that's the only way it could get that linear read and call it good.

On nforce570 SATA, Microsoft default drivers:
#1 Seagate 7200.10 250GB 40GB C drive NTFS 4KB cluster
#2 Seagate 7200.9 160GB 160GB D drive NTFS 8KB cluster

http://evermorex.home.comcast.net/random/diskspeed01.jpg
http://evermorex.home.comcast.net/random/diskspeed02.jpg
http://evermorex.home.comcast.net/random/diskspeed03.jpg
http://evermorex.home.comcast.net/random/diskspeed04.jpg

IDE vs. SATA isn't what's making lower scores. Linear read speed on regular drives and controllers is limited by the physical drive speed and density, not the interface speed. Even with a 16MB cache, that's only a fraction of a second's worth of data transfer from the cache to the controller, limited by the interface speed, the rest is all physical transfer rate. Performance of the controller and the drive electronics also comes into play.
 
Originally posted by: Lord Evermore
And with a linear read score like that, how could it possibly get such a high final score?
Click to expand...

It counts the access time as well. But I agree -- the score is meaningless by itself. It's better when the components are seen.
 
Originally posted by: Madwand1
Originally posted by: Lord Evermore
And with a linear read score like that, how could it possibly get such a high final score?
Click to expand...

It counts the access time as well. But I agree -- the score is meaningless by itself. It's better when the components are seen.
Click to expand...

Great. So you can REALLY QUICKLY start slowly transferring your data. 🙂
 
I don't think this app was designed to test flash.

Flash memory is much slower than a hard disk though. The access time is good obviously but the raw transfer rate is quite subpar. The nice thing about this handy program is it's tiny, requires no installation and gives a result in seconds which can be used to compare other drives, etc.
 
Originally posted by: Minerva
The results are similar with other popular utilities such as ATTO, HD Tach, I/O Meter, COSBI, DriveBench. Sandra gets an i/o error for some reason.
Click to expand...

While your array might get similar results with such tools, (1) these benchmarks aren't equivalent, and some of them have issues with some RAID configurations (2) Roadkil's has issues with some RAID configurations for similar reasons.

Re: (1) ATTO has problems due to using small data sizes and request sizes. HDTach has issues with some RAID arrays. IOMeter is not in the same class. It can of course be very easily misused, and is, but in the hands of people who knows what they're doing, it can give more meaningful results. I'm not familiar with the others.

I'll illustrate (2):

Here's a screenshot of a "real world" copy & paste large file transfer, showing 240 MiB/s ~ 252 MB/s:

http://i89.photobucket.com/albums/k203/Madwand0/Vista-local-copy-explorer.png

Roadkil gives 118 MB/s -- which is way off: http://i89.photobucket.com/albums/k203/Madwand0/Roadkil-nvr046464.png

IOMeter, configured appropriately, gives 248 MiB/s, or 260 MB/s, which corresponds well with the "real world" measurement.

http://i89.photobucket.com/albums/k203/Madwand0/iometer-nvr0464641mr.png

I agree that Roadkil's is a nice little useful program. I like it. It's probably just fine for most ordinary cases. But it's not perfect, etc.

Originally posted by: Minerva
It's the closest thing to solid state storage one can get without spending a small fortune. I like it. 🙂
Click to expand...

You have an uber system that's enviable, and sets some sort of benchmark of achievement itself. I think ordinary people would contest the "without spending a small fortune" part though.
 
The RaptorX is a single drive, I'm running two 3200KS using WinXP disk spanning...essentially single drive. My system is an Athlon X2 4800+, 2 GB of Corsair 3500LL memory, Asus A8N32 Deluxe MB - all optimized by me for max performance and stability.

I also defrag regularly - it really does help. 🙂
 
6851 on Seagate 7200.8 300GB 16MB....

Meh, I'm getting a new HD soon, hopefully it will be a little faster and quieter.
 
I'm not sure how good this program is I ran it and got 2500 then I ran it again and got 7439 same drive Maxtor 250 GB 16MB Cache
 
Originally posted by: EricMartello
Here's what I got:
Western Digital RaptorX - 65,069
Western Digital 3200KS - 42,155
Click to expand...

A single Raptor giving 158 MB/s STR, and a JBOD of normal drives giving 145 MB/s STR?

Something's seriously wrong with these benchmarks.

Try doing real, large (e.g. 8 GB) file transfers among these drives. The drives simply do not physically perform this well by themselves; no single drives in this class do.
 
7107 for C: (first partition on my Samsung 250GB SATA 2 drive), 7509 for D: (2nd partition), 8034 for the whole physical drive.
6634 for E: (Maxtor 250GB SATA 1).
 
Originally posted by: Madwand1
Originally posted by: EricMartello
Here's what I got:
Western Digital RaptorX - 65,069
Western Digital 3200KS - 42,155
Click to expand...

A single Raptor giving 158 MB/s STR, and a JBOD of normal drives giving 145 MB/s STR?

Something's seriously wrong with these benchmarks.

Try doing real, large (e.g. 8 GB) file transfers among these drives. The drives simply do not physically perform this well by themselves; no single drives in this class do.
Click to expand...

Yes it seems that one run per reboot is required. Cache whether hardware or Windows based is skewing the results. When I run it subsequently it goes even higher!

File copy results (especially in Vista) cannot be trusted as a benchmark due to caching. Iometer is showing its age too as the effect of huge hardware cache skews the results. HOWEVER it can be said those seemingly inflated results do translate into real world improvements.

To gauge actual file copy performance try this benchmark.
Commonly referred to as COSBI or Comprehensive Open Source Benchmarking Initiative, this one lets you adjust parameters and source/targets for more thorough analysis of system performance.

It MAY be affected by windows disk cache particularly in newer versions of Windows. I have not really tested them.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top