Mid-high range gaming PC.

[_Neptunno_]

@defaultuser said in Mid-range gaming PC.:

@_neptunno_ you also have raid 0+1 and raid 5.
With these you can have concatenated raid just like in raid 0, but protected against a disk failure.

Salu2.

Si, only I have summarized

[Clipper]

A question, if I physically disassembled the raid 0 and mounted it on ports 2 and 3, does it work or do I have to reinstall the OS?
Another question, I bought a key (legal) for W10 pro, theoretically it's an OEM for a single computer but... I get an activation error.
Any reliable website to get another one or should I go to Amazon (which didn't give me problems before, on the website I bought it from it does give me problems)
Regards

[_Neptunno_]

@clipper in theory nothing should happen, I have messed around several times with Raid 0 and nothing happened to me... although in case of doubt make a backup if you have important data. But I swear that nothing should happen, the problem is if a disk is damaged since you lose half of the information.

About the Key, insist with the seller and if not use Amazon. But the reliable key pages are very well known if you see any video on Youtube about computer topics you will see the same websites.

Regards.

[Clipper]

For unknown reasons the crystaldisk does not "see" the raid 0 and I can't pass the test.
Any ideas?
The W issue was resolved by buying a key for W11 pro.
But before installing it I would like to see the performance of the raid 0 and an m2 with only w10.
I prefer to leave W11 mounted on the raid 0 of 3 m2 (if the result does not convince me I will leave only one m2)

[Sylver]

@clipper It seems that CrystalDisk can't read the S.M.A.R.T. values of a disk array, but it can read them separately.

To see performance and health status with the Raid mounted, you need some specific Raid software, or you can also try with HDTune, which in theory allows you to run tests.

Regards

[Clipper]

In theory, HD tune only allows you to do the read test.
Except in the paid version.
And with that, we are left halfway in serious tests...
Regards

[defaultuser]

@clipper If I remember correctly, it does sequential reading and random operations, with that you may have enough to check if the raid improves more or less, you can extrapolate the writes based on the disparity that the disks already have between writing/reading, that shouldn't change.

Knowing the performance of the disks separately beforehand, hdtune is enough to know how much the raid yields in general.

If you want to do a more thorough analysis, look for a limited time demo or something like that, or if there is any freeware out there.

[jordiqui]

@Clipper you are on the team of Hwot, could you pass a few benchmarks to improve the team's score. If I'm not mistaken, it's @krampak our captain

[Clipper]

That's done.
In a few days I'll upload it.

[Clipper]

Mounted the RAID 0 of three mn2.
The results are...
Curious.
In a while I'll upload the screenshot.
Regards

[Clipper]

greetings

[defaultuser]

@clipper ¿Que discos son?

[Clipper]

@defaultuser crucial P5 pro 500 gs
But it seems to me that it is not profitable...
Regards

[defaultuser]

@clipper said in Mid-range gaming PC.:

crucial P5 pro 500 gs

Most likely it won't.
I would get a good one for the system, and those three I either sell them or I add them to a raid and use it as primary storage.

[Clipper]

@defaultuser what doesn't add up to me is this: https://www.crucial.es/ssd/p5-plus/CT500P5PSSD8/CT20177648
And it is that RAID 0 is below the values that Crucial indicates for a single Mn2.
The only thing that comes to mind is that slot 1 is linked to the CPU and 2 and 3 are not and that creates a bottleneck.
Regards

[defaultuser]

@clipper
@defaultuser said in Mid-high range gaming PC.:

@clipper The first M2 slot is controlled by the cpu, the other two by the chipset.
The chipset communicates with the cpu through a pcie x4 channel, in theory it has plenty of bandwidth, but the response times will always be worse than in slot 1.
A raid0 in the first two slots probably performs poorly, and I think that's what you've set up based on the results.
A raid0 in the chipset should increase the transfer, but we'd have to see how much, and in any case with the speed that those disks already give by themselves....
To sum up, setting up the raid0 in the chipset to have a secondary large disk without paying the prices of "large" disks maybe not, but increasing performance with a raid0 on this platform I don't see it, you'll have to "settle" for the speeds of a single disk, which are already the best.
Now you're not really doing badly, but comparing prices, your two disks sold could even pay directly for the same but 1TB, or you can have double disk and have system on one side and games on the other and things like that, that is, a disk in the first slot and in the chipset another or a raid.

[Clipper]

@defaultuser two disks not...
Three disks in raid 0.
Regards
P.d the first screenshot I posted was a raid 0 of two disks on ports 1 and 2
The last one is three disks on ports 1 2 3.
Regards

[defaultuser]

@clipper If correct, but it doesn't matter for the case.
By putting the chipset into the equation, that is achieved, there are many differences from one way to the other.

[Clipper]

Well, after the fiasco of the RAID 0, I'll leave it on an mn2 for the OS
And I'll look at the raid 0 on ports 2 and 3 to see performance and do tests.

After which, depending on the results, I'll do one thing or another, sell the mn2s or keep them.

Since I already have the laser cutter operational at work, I'll make the parts for the pull/push of the CPU aio.
Initially the design was more complex... But simple is more effective.
I'll leave it in a 5mm methacrylate with the holes for the fans/radiator (they are the same) and a 10mm lid with the holes for the fans.
The assembly will be complex but I think effective.
The top cover will be a "four-petal margarita" system, I think it's the most effective.
In a few weeks, photos
Regards

[Clipper]

Curiously for other reasons (noise, I thought the GPU bracket was touching one of the fans) I removed the side cover.
Results...
I went from almost 100 degrees in the same game to 50 degrees.
We're talking about a tower that brings in air with 6 fans plus one in the back and exhausts it all from the top through the AIO.
It didn't have bad temperatures in the CPU but in the GPU.
Conclusion:
The GPU needs direct air intake.
And 80% of the towers don't have it.
Which looks great with the F.A. down below with a cover... But in terms of cooling it's a chestnut.
So for now I will launch the modular tower project without R.L for the GPU.
Regards