SSD benchmark

[invision2212]

Is there a tool to test out the speed of the SSD?

[Peekay]

Publicado originalmente por PopinFRESH:

Publicado originalmente por Peekay:

The deck does cap both of these drives with its limitations and you can see it in benchmarks but believing marketing specs on SSD's is hilarious. Both of the 256GB and 512GB drives are only benched by decks that can't prove anything. I'll be getting my deck soon and I'm willing to bench it's 512GB SSD in my PC if someone with the other drive is willing to do the same we can get some real numbers on both drives.

So you're suggesting that the marketing of the SSDs is some how lower than what the SSD will actually perform like? You have an odd view on how companies market things.

I'm pretty sure you're trolling at this point but you're free to take one of your 970 Evo Plus NVMe PCIe M.2 1TB's and benchmark it yourself. I'm sure you'll get the advertised

3,500MB/s Seq. Read
3,300MB/s Seq. Write
Lmao.

[PopinFRESH]

Publicado originalmente por Legangsta:

Btw, google OM3PDP3256B-A01
Exact page layout and set-up, writing, numbers etc. as what you copy/paste-spam : https://exe.ua/en/product/p375947/

Look here it's a bit more serious :
https://www.digikey.com/en/products/detail/kingston/OM3PDP3256B-A01/15822181

Read 2400.

2400 > 1700

3x4 performs better than 3x2, no?

BTW Digikey is quoting a generic spec from the general design-in PDF, not the specific SSD but cool story anyway.

I actually got the specs from Arrow which previously had the actual specifications for the different capacities for that model. e.g. OM3PDP3128B-A01, OM3PDP3256B-A01, and OM3PDP3512B-A01. They had them back when reviewers were originally publishing their tear downs and listing various part numbers for the components. Unfortunately it looks like they no longer have it available as its now listed as Obsolete and all that is listed now is a link to the generic design-in SSD PDF [static6.arrow.com] .

Again, the 512GB version will hit 2400MB/s seq read at a high queue depth in a synthetic benchmark. The 256GB version is 1700MB/s seq reads. Neither is going to be sustained above 2GB/s in real world workloads. Feel free to take any of them out of your Steam Deck, slot it into your desktop Windows machine and run PCMark 10 Storage benchmarks on it. You're going to find that none of the 6 drives in question are going to saturate 2GB/s

[PopinFRESH]

Publicado originalmente por Peekay:

Publicado originalmente por PopinFRESH:

So you're suggesting that the marketing of the SSDs is some how lower than what the SSD will actually perform like? You have an odd view on how companies market things.

I'm pretty sure you're trolling at this point but you're free to take one of your 970 Evo Plus NVMe PCIe M.2 1TB's and benchmark it yourself. I'm sure you'll get the advertised

3,500MB/s Seq. Read
3,300MB/s Seq. Write
Lmao.

Yeah I do, if its a clean secondary drive using a synthetic benchmark like CrystalDiskMark that is what I get. But that isn't reflective of real world performance. You are the one who seems to be suggesting that number is some how lower than how an SSD will actually perform.

If you are suggesting the advertised numbers are "higher" than what you'll actually see in real world use then feel free to explain to me how 1700MB/s or less is going to saturate a 2000MB/s link? And the 512GB versions which will hit 2400MB/s sequential reads with high queue depths in synthetic benchmarks aren't going to sustain more than the 2GB/s in real world workloads either.

[Peekay]

Publicado originalmente por PopinFRESH:

Publicado originalmente por Peekay:

I'm pretty sure you're trolling at this point but you're free to take one of your 970 Evo Plus NVMe PCIe M.2 1TB's and benchmark it yourself. I'm sure you'll get the advertised

3,500MB/s Seq. Read
3,300MB/s Seq. Write
Lmao.

Yeah I do, if its a clean secondary drive using a synthetic benchmark like CrystalDiskMark that is what I get. But that isn't reflective of real world performance. You are the one who seems to be suggesting that number is some how lower than how an SSD will actually perform.

If you are suggesting the advertised numbers are "higher" than what you'll actually see in real world use then feel free to explain to me how 1700MB/s or less is going to saturate a 2000MB/s link? And the 512GB versions which will hit 2400MB/s sequential reads with high queue depths in synthetic benchmarks aren't going to sustain more than the 2GB/s in real world workloads either.

It's almost like you understand SSD's need to be tested to see how they really perform.

[PopinFRESH]

Publicado originalmente por Peekay:

Publicado originalmente por PopinFRESH:

Yeah I do, if its a clean secondary drive using a synthetic benchmark like CrystalDiskMark that is what I get. But that isn't reflective of real world performance. You are the one who seems to be suggesting that number is some how lower than how an SSD will actually perform.

If you are suggesting the advertised numbers are "higher" than what you'll actually see in real world use then feel free to explain to me how 1700MB/s or less is going to saturate a 2000MB/s link? And the 512GB versions which will hit 2400MB/s sequential reads with high queue depths in synthetic benchmarks aren't going to sustain more than the 2GB/s in real world workloads either.

It's almost like you understand SSD's need to be tested to see how they really perform.

It's almost like we know they aren't going to perform faster than their specs would suggest. Feel free to pop yours into your desktop and run it through PCMark 10 storage benchmarks to see it not exceed 2GB/s. You seem to think the issue is a CPU bottleneck is "holding back" the SSDs from what they would actually be able to do, which somehow is faster than the expected specifications. You aren't going to see any meaningful difference between running CrystalDiskMark on the Steam Deck vs the SSD in your desktop PC.

[Peekay]

Publicado originalmente por PopinFRESH:

Publicado originalmente por Peekay:

It's almost like you understand SSD's need to be tested to see how they really perform.

It's almost like we know they aren't going to perform faster than their specs would suggest. Feel free to pop yours into your desktop and run it through PCMark 10 storage benchmarks to see it not exceed 2GB/s. You seem to think the issue is a CPU bottleneck is "holding back" the SSDs from what they would actually be able to do, which somehow is faster than the expected specifications. You aren't going to see any meaningful difference between running CrystalDiskMark on the Steam Deck vs the SSD in your desktop PC.

You understand the deck is bottlenecking both SSD's and that's why they're performing the same right? I can't take you seriously.

[PopinFRESH]

Publicado originalmente por Peekay:

Publicado originalmente por PopinFRESH:

It's almost like we know they aren't going to perform faster than their specs would suggest. Feel free to pop yours into your desktop and run it through PCMark 10 storage benchmarks to see it not exceed 2GB/s. You seem to think the issue is a CPU bottleneck is "holding back" the SSDs from what they would actually be able to do, which somehow is faster than the expected specifications. You aren't going to see any meaningful difference between running CrystalDiskMark on the Steam Deck vs the SSD in your desktop PC.

You understand the deck is bottlenecking both SSD's and that's why they're performing the same right? I can't take you seriously.

No it isn't. The single NAND memory chip isn't performant enough for it to matter.

[The Green Nerd]

I just found out that you can download Kdiskmark appimage and just open in. Only up to version 2.3.0. After that you need sudo.
https://github.com/JonMagon/KDiskMark/releases/tag/2.3.0 --> https://github.com/JonMagon/KDiskMark/releases/download/2.3.0/KDiskMark-2.3.0-fio-3.28-x86_64.AppImage

[Spooky]

honestly what the hell, lol... this is driving me nuts. I have run kdiskmark on my 64gb deck in the past, while drunk and trying to figure out what the real diff between micro sd and emmc and sata drives and sata ssds and m.2 etc, lol. I vaguely remember the internal clocking in at around 500mB/s ? jeje.. faster than I expected tbh, to put into perspective micro sd cards max out at around 120mb/s, similar to sata3 controllers for spinny hdds hehe. but I think close to sata3 ssd speeds iirc., not awful in any way. and still 5 times faster than most micro sd cards.

I installed a gen4x4 wd drive the other day mainly because I was literally tired of trying to keep everything on sd cards and then realising that heh.. swap cards and the symlinked everything goes to hell etc... idk... there are solutions but all of it is such a weird headache that literally was never advertized on the tin. I had no issues with 64gig internal, why would I... sd cards exist... hehehe... then slowly notice that what tiny portion of the 64gig internal you have left gets swallowed for every game that you install.. sigh.

hayy, I figured it out... emm... it is that version above, 2.3.0 but after downloading the appimage right click and set it as an executable.

and if anyone is curious I guess my deck was maybe one of the early ones taht they decided to shaft users by saying screw it, gen3 is more than anyone needs... hehehe... like obviously yest, 3.6gb/s compared to every drive they shipped decks with from the drive that I bought is badass... but... at the same time... when I clicked buy, within hours of it being launched the product description literally said gen4x4 ... lol... even at the time a year later when I was asked to pay the full amount it said gen4x4. You owe me 2.5gb/s valve, lol, rofl... not a joke... sigh.

[Spooky]

half a joke.. I still feel a bit miffed that I am not getting the full speed, but also hehe, I was a bit worried that it actually running at 4x4 speed would create too much heat and maybe not be ideal. at this speed from what I can tell the drive draws less power than any of the stock steam deck m.2s that are released... not by much, like talking probably, hmm.. 0.1-0.3 watts lower ? hehehe...

It was the WD SN740 from ali that I nabbed, at the time I was tempted to get their gen3 cards that were slower rated, just in case.. but lol... emm.. their max watt draws at I guess slower than this is running were more power intensive and well, more watts = heat as a byproduct...

I seem disappointed but lol, honestly.. sigh... never seen even 3.6gb a sec on a drive before hahaha.. shame the 512 I got is almost full with I think me only installing titanfall 2 on it .. lol.... it is ok, I will survive ;)