Sandisk SDSA6AM-008G:

Mostly fresh TLC (I think) based drive which was scavenged from a printer. Fast on small blocks, but may be problematic to use without TRIM in the long run.

Toshiba MLC based SSD TSSM1TA24M008G (manufacturer unknown):

Fairly slow NAND memory, which was mostly used in el cheapo USB sticks. Still, easily can compete with an old SLC drive.

InnoDisk FiD 2.5 SATA10000:

Very fast 8Gb SLC based drive, at least when it comes to sequental writes. For some reason does not like lots of random small blocks in CrystalDiskMark. Shouldn't be a problem in real scenario though.

Transcent TS8GSSD25S-S:

Almost identical to InnoDisk SATA10000.

Transcend 1Gb 40-pin IDE Flash module (circa 2011?):

Strange behavior on a few small block writes (ATTO), but overall not bad for 1Gb SLC (maybe) drive.

InnoDisk EDC4000:

Veeery slow on small blocks (CF cards levels of bad), but starting from 4k blocks speeds up quite decently. Most likely common problem of large PATA DOMs. SLC based.

I use 2.5 pata to msata, those cheap laptop pata enclosures they are like 5 dollars, + a laptop to desktop ide converter they are like 2 dollars. Then I just put whatever msata i wany inside and just plug and play to ide, it has worked on everything i ever tried it on. and you can put any size ssd inside u want.

I usually prefer more robust old SLC drives, which are perfectly capable to work in PIO modes even without an adapter. They are extremely cheap these days too, if you know where to look.
Also I should note that even 4Gb Apacer feels quite comfortable and fast to use under WinXP.

P.S.
If someone else interested to post their benchmarks, feel free to do so. It would be quite interesting to see how different IDE DOMs behave.

I'm putting together a spreadsheet of my results.

I have been benchmarking with ATTO under Win98SE.
[*] Is it worth looking at a DOS benchmark?
[*] I'd like to eventually bench mark something real-world I/O bound like boot time or scripted install time.

I selected 2 motherboards to bench with right now:
[*] Gateway BAT4IP3 (420EX) w/ Amd 5x86-WB 133+ (Bios supports 2GB drives, but turns to stone if it sees > 2GB, EZ overlay works with 2GB config if I keep the partitions under 8GB, gotta get a better bios )
[*] P5I430TX TITANIUM IB+ w/ Cyrix MII 300+ ( Rev J bios supports up to 128GB, love it)
[*] Loving scripted install under win98se.
[*] Keeping those Roms shadowed
I can add results from a 386sx20, 486dx33VLB and a Celeron 800, but my testing desk only has space for 2 setups right now and getting those things stable takes time.

I'm testing where I can with these IDE controllers --
[*]Barebones ISA IDE that came with a CD ROM - I call it "the Gimp". Might be 0.75" tall. More like a shim than a card.
[*]Paradise 100TX2 (Doesn't detect drives on my 486 motherboards)
[*]Matrox 150 ( Looks like a Paradise 150Tx2 but with different bios, locks up on boot with my 486 motherboards)

I have a couple IDE controllers I hope to get to but --
[*]XT IDE - Have to take some time to get the jumpers & bios right so it works on AT class
[*]VIA Raid - has no bios / software won't mount single drives / but has great ATA matrix info & Smart reporting
[*]Adaptect Ultra 66 -( ASH-133 w/ SIIG Chipset ) - I have not had this work anywhere yet. I should try it in the 440BX board or return it soon
[*]Promise ATA33Ultra - Should arrive this week along with the last DOM
[*]SIIG SATA 150 - Currently in a working 440BX box but it is SATA only and i'm not testing that many Sata devices

I have a pile of drives to benchmark, from a 1993 Quantum Prodrive (Came back to life after a lot of prodding) to a Crucial M4 SSD
[*]Have not been able to get a 44pin laptop module to recognize yet, pins are not labeled, may have fried it by getting pin 1 wrong. At least it was really cheap.
[*]Had to get out a freedos boot disk to partition the 256GB SSD. Win98SE FDISK did not like it. 0% .. 0% .. 0% .... 20 mins later .. 1%

I'm collecting these data points:
[*]Motherboard/CPU
[*]Controller
[*]Drive
[*]ATA / UDMA Levels
[*]Read/write ATTO timings

Once I'm don't bench marking:
[*]I'll play with ATA secure erase to see if I can show that it helps DOMs/SDs or CFs.
[*]I'll crack open the DOMs to get controller info

The biggest things I've noted so far are:
[*]Compatibility is a lot harder than I expected for new drive & a 15 year old controllers in 25 year old motherboards
[*]Probably the most important thing for good performance is getting the proper ATA/UDMA/Multiword negotiation. This can be tricky. I can give hints what I want in the bios, but what gets negotiated can be different.
[*]All of the SD cards I've used have been limited by the speed of the IDE/SD adapter (Single word UDMA 5) . Need to dig up some older SD to find something that is SD limited
[*]An old 6.4 GB Quantum Fireball EX holds its own against the small DOMS, IDE/SD and CF adapters because of the 512K onboard cache. Wipes the floor with them on large transfers because it negotiates functional multi-word DMA.

This site had lots of good info about the limits for the different chips used in the adapters.

https://thechipcollective.com/posts/cynical/ewaste/

I found it when I was looking at the SATA-SD adapters on ebay and noticed that it was a two chip implementation using chips I already have. Here is what they had to say about the performance of the two chip device:

Consulting the datasheets reveal how it works but not why it’s so slow. The JM20330 is a SATA to PATA bridge, and the FC1307A is a CompactFlash to SD adapter. Since CompactFlash cards can run in ATA mode and are essentially tiny PATA SSDs, the two adapters can be chained. This occurs within the board, hence the large parallel bus visible between the chips. The SD card is converted to PATA which is converted to SATA which does indeed create a SATA SSD. It’s extremely shit and outperformed by an X-25M, but it is technically a SATA SSD.

The JM20330 was released in 2003 if we go by the datasheet’s publication date. SATA was only three years old, and lots of PATA drives were still floating around, so it made sense to produce a bridge that allowed the use of old drives in new systems. It supports ATA/ATAPI-7 and UDMA 6, giving it a 133 MB/s parallel ATA bus and a 150 MB/s serial ATA bus. That shouldn’t be the cause of any slowdowns4.

The FC1307A appears to support UDMA 6 on the host side, the same as the JM20330 (though the datasheet only claims ATA/ATAPI-6 compatibility). On the device side it supports SDHC 2.0, which allows cards to run at 25 MB/s. This still doesn’t explain the slowdown because that’s six times the transfer rate benchmarked5.

And individually, neither chip is slow. A series of tests indicates that the FC1307A is able to sustain 24 MB/s6, and although there are no real benchmarks of the JM20330 by name, there is a claim that it can hit 30 MB/s with a hard disk drive7. It’s definitely the combination of the two chips that’s causing the issue, and JMicron’s apparent inability to make good chips involving SATA doesn’t help.

Avoid this. Never buy this. This is a terrible, terrible piece of e-waste. It never should have been made, and it was made to be tossed into a landfill.

I was seeing if there was any easy way to get around the 25MBs bottleneck on the FC1307A chip commonly used in IDE-SD adapters.

I tried connecting an IDE to USB adapter with a USB to flash adapter.
It worked in DOS. Win98se was not happy with it though. It didn't get past the loading screen, even though the color cycling at the bottom of the screen kept running and num lock still worked.

Those IDE flash memory are compatible with win98 out of box?
Or just has the same issues of ssd?

douglar wrote on 2020-03-17, 16:46:

This site had lots of good info about the limits for the different chips used in the adapters. […]
Show full quote

This site had lots of good info about the limits for the different chips used in the adapters.

https://thechipcollective.com/posts/cynical/ewaste/

I found it when I was looking at the SATA-SD adapters on ebay and noticed that it was a two chip implementation using chips I already have. Here is what they had to say about the performance of the two chip device:

Consulting the datasheets reveal how it works but not why it’s so slow. The JM20330 is a SATA to PATA bridge, and the FC1307A is a CompactFlash to SD adapter. Since CompactFlash cards can run in ATA mode and are essentially tiny PATA SSDs, the two adapters can be chained. This occurs within the board, hence the large parallel bus visible between the chips. The SD card is converted to PATA which is converted to SATA which does indeed create a SATA SSD. It’s extremely shit and outperformed by an X-25M, but it is technically a SATA SSD.

The JM20330 was released in 2003 if we go by the datasheet’s publication date. SATA was only three years old, and lots of PATA drives were still floating around, so it made sense to produce a bridge that allowed the use of old drives in new systems. It supports ATA/ATAPI-7 and UDMA 6, giving it a 133 MB/s parallel ATA bus and a 150 MB/s serial ATA bus. That shouldn’t be the cause of any slowdowns4.

The FC1307A appears to support UDMA 6 on the host side, the same as the JM20330 (though the datasheet only claims ATA/ATAPI-6 compatibility). On the device side it supports SDHC 2.0, which allows cards to run at 25 MB/s. This still doesn’t explain the slowdown because that’s six times the transfer rate benchmarked5.

And individually, neither chip is slow. A series of tests indicates that the FC1307A is able to sustain 24 MB/s6, and although there are no real benchmarks of the JM20330 by name, there is a claim that it can hit 30 MB/s with a hard disk drive7. It’s definitely the combination of the two chips that’s causing the issue, and JMicron’s apparent inability to make good chips involving SATA doesn’t help.

Avoid this. Never buy this. This is a terrible, terrible piece of e-waste. It never should have been made, and it was made to be tossed into a landfill.

SATA to SD.jpg

I can't yet speak for the JM20330 (got some on order), but its little brother, the Jmicron JMD330 is no slouch .
This is in P3 1400MHz 512K system with an i815ep/ICH2 ATA66 controller and Jmicron JMD330 hooked up to a Samsung 860EVO (so the drive is not a bottleneck) and running Windows 98 SE.
It's even faster on a Promise Ultra133 TX2 (90MB/sec read in HDTACH 2.61, did not try ATTO)

While some people have gotten a lot of free time from the quarantine, I have not. Really sidelined my storage benchmarking project.

One thing I've learned is that some of those cheapo counterfeit SD cards you can get from the grey market have some truly atrocious throughput and uneven performance. I got a knock off Samsung 128GB micro sd that maxes out at a 7MB/s transfer rate and it rarely sustains that for any length of time. Wasn't worth $15. It's a bottleneck even on a 486.

While I like the thoroughness of ATTO for bench marking, it's tough to compare because there are so many data points.

Has anyone used MS-DOS Disk performance tester, by James Pearce ? https://www.lo-tech.co.uk/wiki/DOS_Disk_Tester

I like the 4 tests that it has:

• 32K Sequential Read (KB/s)
• 32K Sequential Write (KB/s)
• 8K Random (IOPS) - 70% read, 8K aligned
• 512-byte (single sector) random read IOPS, sector aligned

douglar wrote on 2020-05-20, 15:42:

While some people have gotten a lot of free time from the quarantine, I have not. Really sidelined my storage benchmarking proj […]
Show full quote

While some people have gotten a lot of free time from the quarantine, I have not. Really sidelined my storage benchmarking project.

One thing I've learned is that some of those cheapo counterfeit SD cards you can get from the grey market have some truly atrocious throughput and uneven performance. I got a knock off Samsung 128GB micro sd that maxes out at a 7MB/s transfer rate and it rarely sustains that for any length of time. Wasn't worth $15. It's a bottleneck even on a 486.

While I like the thoroughness of ATTO for bench marking, it's tough to compare because there are so many data points.

Has anyone used MS-DOS Disk performance tester, by James Pearce ? https://www.lo-tech.co.uk/wiki/DOS_Disk_Tester

I like the 4 tests that it has:

• 32K Sequential Read (KB/s)
• 32K Sequential Write (KB/s)
• 8K Random (IOPS) - 70% read, 8K aligned
• 512-byte (single sector) random read IOPS, sector aligned

Lack of DMA support for many IDE interfaces under DOS (at least without using UDMA.SYS and friends) means PIO mode is a common bottleneck for most devices . To be representative of actual max device performance, DMA mode would have to be enabled manually where not done automatically.

This is, again, in a P3 1400MHz 512K system with an i815ep/ICH2 ATA66 controller and Jmicron JM20330 this time hooked up to a Samsung 860EVO (so the drive is not a bottleneck) and running Windows 98 SE.

darry wrote on 2020-05-21, 22:05:

Lack of DMA support for many IDE interfaces under DOS (at least without using UDMA.SYS and friends) means PIO mode is a common bottleneck for most devices . To be representative of actual max device performance, DMA mode would have to be enabled manually where not done automatically.

I finally got a chance to play around with disktest.exe today.

It's fine & concise for measuring DOS / BIOS access, but those types of access are often missing out on UDMA, multiword transfers, etc, as you mention.

The numbers that come out when running in a dos session in Windows seemed a little inconsistent. I need to test a few more setups to see if it is useful.

darry wrote on 2020-05-21, 22:08:

This is, again, in a P3 1400MHz 512K system with an i815ep/ICH2 ATA66 controller

ICH2 is ATA100 though.

The Serpent Rider wrote on 2020-09-24, 02:37:

darry wrote on 2020-05-21, 22:08:

This is, again, in a P3 1400MHz 512K system with an i815ep/ICH2 ATA66 controller

ICH2 is ATA100 though.

You are absolutely right !

After installing the drivers from the installation CD of my industrial board and getting results consistent with ATA66 , I never bothered checking .
The results were as follows :

After reading your post and checking the datasheet, I downloaded and ran IATA_CD.EXE which updated my driver. I then ran a test which gave the following strange results . I also tried running a test with HDTACH, but that crashed the system with the new driver, so I rolled back .

EDIT: I am using the same Jmicron IDE to SATA adapter that I was using with a Promise ULTRA133 TX2 and that was giving me about 90MB/second .