Poor Benchmark Performance, Fine Real World Performance

[PrinzyXX]

Now I think I may be going crazy here but get this. My brother was curious about my raid 0 performance (don't start with why I chose to use raid 0, I have a backup of the raid in case of raid failure) on my two 2TB FireCuda hard drives from Seagate, and when I ran the crystaldisk benchmarks they show abysmal performance, but that's not translated into real world performance because my games dont take that long to load up. Could someone please ease my madness here? There's no issue as far as I can tell, but these numbers are driving me crazy. The data sheet for the drives say an avg read performance of 156 MB/s and avg write performance of 190 MB/s and since they are in raid 0 that makes a theoretical 312 MB/s read and 380 MB/s write, unless my logic is flawed there. I've attached benchmark results from crystaldisk doing a single pass of read/write tests of a 100MB file, both sequential and 4K random.

[RadiatingLight]

undo the raid 0, it will probably return to normal performance. something about the RAID is messing up read/write speeds.

[PrinzyXX]

Just now, RadiatingLight said:

undo the raid 0, it will probably return to normal performance. something about the RAID is messing up read/write speeds.

you think so? I mean its software raid so I know it isnt the best, but, I'm boggled by the fact that whenever I did a file transfer test, windows task manager is reporting the full speeds I'm expecting from the raid.

[RadiatingLight]

1 minute ago, MysticalRainXIV said:

you think so? I mean its software raid so I know it isnt the best, but, I'm boggled by the fact that whenever I did a file transfer test, windows task manager is reporting the full speeds I'm expecting from the raid.

Dunno.

Maybe use another benchmark

 

[Mira Yurizaki]

I'd say if your real-world performance appears normal, then chalk it up to a fluke and not worry about what the benchmark reports.

[PrinzyXX]

Just now, RadiatingLight said:

Dunno.

Maybe use another benchmark

 

Ive used atto, crystaldisk, and even the ones built into the software form my 850 evo and pirifiorms defraggler. It's just..odd and it bothers me. 

 

[PrinzyXX]

Just now, M.Yurizaki said:

I'd say if your real-world performance appears normal, then chalk it up to a fluke and not worry about what the benchmark reports.

I agree, but its still an odd issue, because benchmarks before have reported the real world performance I expect. It recently started happening.

 

[Mira Yurizaki]

3 minutes ago, MysticalRainXIV said:

I agree, but its still an odd issue, because benchmarks before have reported the real world performance I expect. It recently started happening.

 

Well, CrystalDiskMark isn't anything special. It's a GUI front-end of a Windows CLI tool.

 

So I guess you could blame Microsoft.

[PrinzyXX]

Just now, M.Yurizaki said:

Well, CrystalDiskMark isn't anything special. It's a GUI front-end of a Windows CLI tool.

 

So I guess you could blame Microsoft.

is this the same for atto? I mean yeah I dont mind blaming microsoft lol

[Mira Yurizaki]

Just now, MysticalRainXIV said:

is this the same for atto? I mean yeah I dont mind blaming microsoft lol

I dunno about ATTO.

[PrinzyXX]

Just now, M.Yurizaki said:

I dunno about ATTO.

Well, I'll keep digging into it and listening for and ideas. I'm not about to break an array for stupid benchmark results, but I question the health of the drives, even though S.M.A.R.T reports everything as normal