VOGONS


Reply 20 of 45, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
Solo761 wrote on 2022-06-24, 08:00:

Issue here is lack od Win9x drivers for Promise SATA300 TX4 card, not maxing out speed.

90 MB/s I get on VT6421A is more than enough for P233 MMX and Win9x. But using Promise SATA300 TX4 card is different matter because of lack of Win9x drivers. You can install ones from Promise website, although they're Win2k/XP, but there won't be any improvement because it just "names" it, it doesn't actually install any drivers, so it ends up working in compatibility mode which is slow.

My numbers are in compatibility mode? Not saying you are wrong, I’m asking.

I haven’t been able to replicate the low numbers you are showing.

I’m looking in device manager, and you’re 100% right. It says no driver files are required for this device. But gives a date of 05/16/2007 and says not signed.

I guess the real question is, given the numbers I posted, how much of a speed bump would I get by going to a different card?

There is a section for command line parameters. Could we use this to set up the device better?

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 21 of 45, by Solo761

User metadata
Rank Member
Rank
Member

AKA DOS mode, or how ever it's called. No advanced features that would result in better speed if OS had drivers to talk "native" language.

Maybe it's combination of SSD and the card. So far I didn't manage to get more than 10 MB/s without drivers. If I install XP and then drivers it goes over 100 MB/s.

When I reinstall win98 today I can take a screenshot of same computer with Via VT6421A card, numbers are way higher.

I doubt any command parameters would work because there aren't any drivers to interpret them.

Reply 22 of 45, by Sphere478

User metadata
Rank l33t++
Rank
l33t++

well, so what's the fastest 9x sata card then?

So if this is a dos mode driver, is there a better dos sata driver we could be using?

BTW,

Updated OP with the drivers that I have. If you all have a version that I don't, please upload.

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 23 of 45, by Riikcakirds

User metadata
Rank Member
Rank
Member
Solo761 wrote on 2022-06-24, 10:27:
AKA DOS mode, or how ever it's called. No advanced features that would result in better speed if OS had drivers to talk "native" […]
Show full quote

AKA DOS mode, or how ever it's called. No advanced features that would result in better speed if OS had drivers to talk "native" language.

Maybe it's combination of SSD and the card. So far I didn't manage to get more than 10 MB/s without drivers. If I install XP and then drivers it goes over 100 MB/s.

When I reinstall win98 today I can take a screenshot of same computer with Via VT6421A card, numbers are way higher.

I doubt any command parameters would work because there aren't any drivers to interpret them.

The Tx4 supports udma 133 via it's own bios extension (and udma33,66,100) in DOS. If you run the MEM command in realmode dos you will have 633KB free instead of 640KB as 7KB is used by the loaded bios extension.
Benchmarking in Speedsys on an ancient Pentium 75, Intel 430fx motherboard from 1995 I get 83MB/s using a 128GB SSD attached to the tx4. It's just limited by the PCI bus bandwidth and 32bit PCI slot, not the P75.
Using the motherboards own IDE controller I get around 8MB/s in DOS as it's limited to PIO mode 4.

Win98 in compatibility mode should default to using the TX4's own bios and then whatever UDMA mode the bios detects the attached drive being compatible with. All SSDs should be detect as UDMA133.
The only problem I can think of is unlike realmode DOS, Win9x is loading some useless driver (ESDI_506.PDR ? ) that is intercepting the cards bios calls and limiting it to PIO mode. Remove that and you should get 80MB/s + the same speed the card does in realmode DOS.

Reply 24 of 45, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
Riikcakirds wrote on 2022-06-24, 13:40:
The Tx4 supports udma 133 via it's own bios extension (and udma33,66,100) in DOS. If you run the MEM command in realmode dos yo […]
Show full quote
Solo761 wrote on 2022-06-24, 10:27:
AKA DOS mode, or how ever it's called. No advanced features that would result in better speed if OS had drivers to talk "native" […]
Show full quote

AKA DOS mode, or how ever it's called. No advanced features that would result in better speed if OS had drivers to talk "native" language.

Maybe it's combination of SSD and the card. So far I didn't manage to get more than 10 MB/s without drivers. If I install XP and then drivers it goes over 100 MB/s.

When I reinstall win98 today I can take a screenshot of same computer with Via VT6421A card, numbers are way higher.

I doubt any command parameters would work because there aren't any drivers to interpret them.

The Tx4 supports udma 133 via it's own bios extension (and udma33,66,100) in DOS. If you run the MEM command in realmode dos you will have 633KB free instead of 640KB as 7KB is used by the loaded bios extension.
Benchmarking in Speedsys on an ancient Pentium 75, Intel 430fx motherboard from 1995 I get 83MB/s using a 128GB SSD attached to the tx4. It's just limited by the PCI bus bandwidth and 32bit PCI slot, not the P75.
Using the motherboards own IDE controller I get around 8MB/s in DOS as it's limited to PIO mode 4.

Win98 in compatibility mode should default to using the TX4's own bios and then whatever UDMA mode the bios detects the attached drive being compatible with. All SSDs should be detect as UDMA133.
The only problem I can think of is unlike realmode DOS, Win9x is loading some useless driver (ESDI_506.PDR ? ) that is intercepting the cards bios calls and limiting it to PIO mode. Remove that and you should get 80MB/s + the same speed the card does in realmode DOS.

So you benched in dos? And got 83mb/sec? On a pentium 75? Wow very nice.

I’m a little rusty on the whole real mode/protected mode thing, last time I encountered it was in A+ cert in the early 2000’s 🤣.

So we can’t load a real mode driver or something with autoexec.bat or config.sys or whatever on windows boot? And have it set the card up right?

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 25 of 45, by Solo761

User metadata
Rank Member
Rank
Member
Riikcakirds wrote on 2022-06-24, 13:40:
The Tx4 supports udma 133 via it's own bios extension (and udma33,66,100) in DOS. If you run the MEM command in realmode dos yo […]
Show full quote

The Tx4 supports udma 133 via it's own bios extension (and udma33,66,100) in DOS. If you run the MEM command in realmode dos you will have 633KB free instead of 640KB as 7KB is used by the loaded bios extension.
Benchmarking in Speedsys on an ancient Pentium 75, Intel 430fx motherboard from 1995 I get 83MB/s using a 128GB SSD attached to the tx4. It's just limited by the PCI bus bandwidth and 32bit PCI slot, not the P75.
Using the motherboards own IDE controller I get around 8MB/s in DOS as it's limited to PIO mode 4.

Win98 in compatibility mode should default to using the TX4's own bios and then whatever UDMA mode the bios detects the attached drive being compatible with. All SSDs should be detect as UDMA133.
The only problem I can think of is unlike realmode DOS, Win9x is loading some useless driver (ESDI_506.PDR ? ) that is intercepting the cards bios calls and limiting it to PIO mode. Remove that and you should get 80MB/s + the same speed the card does in realmode DOS.

Just to be clear, are we talking about SATA150 or SATA300 TX4? There is TX4 card in both SATA I and SATA II variant.

Reply 26 of 45, by cyclone3d

User metadata
Rank l33t++
Rank
l33t++

Don't forget the SATA II-150 cards.

I did track down a few SATA II cards that have Win9x driver support ( not Promise )but they are going to be limited by the PCI bus unless they are put into a system that runs the PCI at 66Mhz... I might try them in a system using an AGP to PCI adapter to see how they perform.

Yamaha modified setupds and drivers
Yamaha XG repository
YMF7x4 Guide
Aopen AW744L II SB-LINK

Reply 27 of 45, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
cyclone3d wrote on 2022-06-24, 23:38:

Don't forget the SATA II-150 cards.

I did track down a few SATA II cards that have Win9x driver support ( not Promise )but they are going to be limited by the PCI bus unless they are put into a system that runs the PCI at 66Mhz... I might try them in a system using an AGP to PCI adapter to see how they perform.

What are the models of these sata II 9x cards?

That would be pretty cool. Use the agp for the storage, and put like a voodoo 5500 in the pci 🤣

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 28 of 45, by aaron158

User metadata
Rank Member
Rank
Member
Sphere478 wrote on 2022-06-25, 00:16:
cyclone3d wrote on 2022-06-24, 23:38:

Don't forget the SATA II-150 cards.

I did track down a few SATA II cards that have Win9x driver support ( not Promise )but they are going to be limited by the PCI bus unless they are put into a system that runs the PCI at 66Mhz... I might try them in a system using an AGP to PCI adapter to see how they perform.

What are the models of these sata II 9x cards?

That would be pretty cool. Use the agp for the storage, and put like a voodoo 5500 in the pci 🤣

there is the sata 150 tx2 and tx4. some versions have the ide connector and some don't. there is also a re branded one with the maxtor name brand on it but its the exact same card as the promise one.

pretty much all other promise cards sata II 150/300 don't have 9x drivers at least not offered by promise them self's.

basically a sata II card would be pointless since the pci bus even on a p3/p4 system seems to bottle neck at around 90-95mb/s so even a Sata 1 card isn't going to be able to do the full 150mb/s

Reply 30 of 45, by Repo Man11

User metadata
Rank l33t
Rank
l33t
aaron158 wrote on 2022-06-25, 00:30:
there is the sata 150 tx2 and tx4. some versions have the ide connector and some don't. there is also a re branded one with the […]
Show full quote
Sphere478 wrote on 2022-06-25, 00:16:
cyclone3d wrote on 2022-06-24, 23:38:

Don't forget the SATA II-150 cards.

I did track down a few SATA II cards that have Win9x driver support ( not Promise )but they are going to be limited by the PCI bus unless they are put into a system that runs the PCI at 66Mhz... I might try them in a system using an AGP to PCI adapter to see how they perform.

What are the models of these sata II 9x cards?

That would be pretty cool. Use the agp for the storage, and put like a voodoo 5500 in the pci 🤣

there is the sata 150 tx2 and tx4. some versions have the ide connector and some don't. there is also a re branded one with the maxtor name brand on it but its the exact same card as the promise one.

pretty much all other promise cards sata II 150/300 don't have 9x drivers at least not offered by promise them self's.

basically a sata II card would be pointless since the pci bus even on a p3/p4 system seems to bottle neck at around 90-95mb/s so even a Sata 1 card isn't going to be able to do the full 150mb/s

The Sil3114 cards have a well deserved reputation for being difficult but I have one that works with my TXP4, and the difference in speed between it and the Promise card isn't much. In both cases, the speed delivered (as measured by Atto) is better than the IDE port of my 8KRA+ with its KT600 chipset.

A lot of times when you first start out on a project you think, This is never going to be finished. But then it is, and you think, Wow, it wasn't even worth it. - Jack Handey

Reply 31 of 45, by zapbuzz

User metadata
Rank Oldbie
Rank
Oldbie

I think FSB benching from 66 to 133 mhz would show significant improvement in operation of SATA II cards due to the bandwidth. Also within the same bandwidth we have spare for multi tasks limiting bottlenecks in real world applications such as file serving or even multimedia/gaming.

Reply 32 of 45, by Repo Man11

User metadata
Rank l33t
Rank
l33t

When I tried the S150 on my PCChips M930 the result showed that the onboard IDE was just as fast as the Promise card so there was no point in using it with this system. This is with a 2.8/512/533 P4, SiS 645, DDR @ 133 MHz, Win98SE and a basic Silicon Power SSD. No SATA card can exceed the limit imposed by the rest of the system.

A lot of times when you first start out on a project you think, This is never going to be finished. But then it is, and you think, Wow, it wasn't even worth it. - Jack Handey

Reply 33 of 45, by Riikcakirds

User metadata
Rank Member
Rank
Member
Solo761 wrote on 2022-06-24, 22:25:

Just to be clear, are we talking about SATA150 or SATA300 TX4? There is TX4 card in both SATA I and SATA II variant.

The card is a SATA300 TX4, picture below.

The attachment tx4by0.jpg is no longer available

This card is very fast in dos, realmode or protected mode (with emm386 loaded). Benchmarks at 83-87MB/s on a Pentium 75 with SSD attached in DOS7.1. No drivers like are loaded in config.sys and no tsrs in autoexec.bat

Win9x should do the same speeds in compatibilty mode (unless something is intercepting the calls to the bios on the card).

Reply 34 of 45, by Solo761

User metadata
Rank Member
Rank
Member

Here are the numbers of Via VT6421A card on already mentioned P233MMX + TX chipset system.

Card has three SATA ports and one IDE port, it's SATA I, up to 150 Mbit.

The attachment vt6421a.PNG is no longer available
Riikcakirds wrote on 2022-06-25, 14:40:

The card is a SATA300 TX4, picture below.

This card is very fast in dos, realmode or protected mode (with emm386 loaded). Benchmarks at 83-87MB/s on a Pentium 75 with SSD attached in DOS7.1. No drivers like are loaded in config.sys and no tsrs in autoexec.bat

Win9x should do the same speeds in compatibilty mode (unless something is intercepting the calls to the bios on the card).

My card looks the same. Didn't try DOS benchmarks, only windows and those are the numbers I get. Windows are Win98SE, without any hacks or twists. Only VGA, LAN and audio drivers. No games or anything, built it few weeks ago.

Reply 35 of 45, by Sphere478

User metadata
Rank l33t++
Rank
l33t++
Riikcakirds wrote on 2022-06-25, 14:40:
The card is a SATA300 TX4, picture below. […]
Show full quote
Solo761 wrote on 2022-06-24, 22:25:

Just to be clear, are we talking about SATA150 or SATA300 TX4? There is TX4 card in both SATA I and SATA II variant.

The card is a SATA300 TX4, picture below.

tx4by0.jpg

This card is very fast in dos, realmode or protected mode (with emm386 loaded). Benchmarks at 83-87MB/s on a Pentium 75 with SSD attached in DOS7.1. No drivers like are loaded in config.sys and no tsrs in autoexec.bat

Win9x should do the same speeds in compatibilty mode (unless something is intercepting the calls to the bios on the card).

Yep that’s my card sata II 300

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 36 of 45, by Mr.Hunt

User metadata
Rank Newbie
Rank
Newbie

Hello! Please consult me. I want to use ssd on old p3v4x, and I think what pci sata controller I must to buy. I think Promise Fasttrack tx4310, I want to use on dosmode and on windows ME (98), as i understand he haven't drivers for 98(me) windows. As I understand, this adapter the same like tx4, but they have chip Promise PDC40719-GP (on tx4 PDC40718-GP). For normal speed i need drivers or not ?

Reply 37 of 45, by MarmotaArmy

User metadata
Rank Newbie
Rank
Newbie

Hi , I have the same card (TX4 Sata II 300)

The driver installs in windows 98... But at least for me is doing nothing. All hard drives are running in MS DOS compatibility mode according to system properties, so I ran ATTO and showed 120MB/s on reading for a cheap Kingdian 32gb ssd which is fine. But I knew something wasnt right because the system as a whole wasn't faster.
So I ran a command window and I typed this.

Copy bigfilename.XXX NUL (I used a 1.7gb rar file which has some backups) (NUL means read only doesn't write)
and timed it to test how much it took
AT 120MB ....lets say 100 it should take ... 20... 30 secs tops
It took like 6 mins... when I divided the size by time I got like 6MB/s , which is PIO limit.

I installed the card on a winxp system and in device manager along the scsi controller a new IDE controller apears , which was not the case with windows 98

Anyone has tried with some of the rloew patches o sweetlow?

Reply 38 of 45, by Sphere478

User metadata
Rank l33t++
Rank
l33t++

Was talking about this with AI earlier and it recommended something interesting.

Has anyone tried this? Can someone try it? (My rig is packed away)

“Method 1: The FastTrak TX4310 Driver Swap (Recommended)
The SATA300 TX4 shares almost identical internal controller logic with Promise's hardware RAID equivalent of the era, the FastTrak TX4310 (which utilizes the PDC40719 chip). Promise did officially release a fully working Windows 98/ME driver for the FastTrak TX4310.
By downloading that legacy package, you can manually point Windows ME to it:
1. Grab the official Promise FastTrak TX4310 Windows 9x/ME driver package from legacy driver mirrors or The Retro Web's Promise Driver Archives.
2. Boot Windows ME, right-click My Computer -> Properties -> Device Manager.
3. Locate your SATA300 TX4 (which is likely showing a yellow exclamation mark or standard PCI storage listing).
4. Click Update Driver -> Specify a Location -> Display a list of all drivers -> Have Disk.
5. Point it directly to the folder containing the FastTrak TX4310 ulsata2.inf and ulsata2.mpd files.
6. Force-select the FastTrak TX4310 from the list when Windows warns you about compatibility.
Because the underlying registers on the PDC40718 chip match the driver's instructions, this drops the native ulsata2.mpd 32-bit miniport driver straight into your Windows ME I/O subsystem, unlocking full burst performance and unthrottling the PCI bus”

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 39 of 45, by CrustyCeleron

User metadata
Rank Newbie
Rank
Newbie
Sphere478 wrote on 2026-10-04, 19:13:

Was talking about this with AI earlier and it recommended something interesting.

Has anyone tried this? Can someone try it? (My rig is packed away)

“Method 1: The FastTrak TX4310 Driver Swap (Recommended)

I have a TX4 active in which to test, albeit only with INT13h.

Unfortunately there is no TX4310 Win9x driver afaik 🙁. I don't see one for any of the same PDC chip (or its generation, either).