In the end I kept the ECS, last week I received the DX version of that mb: http://www.ecs.com.tw/ECSWebSite/Product/Prod … 78-LL-Intel-RR-

Unlucky I have far more issues to make it works properly. With the latest bios version they add the option to enable the hyperthreading, but when I do so the system reboots and after the screen is full of green lines, the only thing I can do is to reset the cmos (I saw another user experiencing this problem in another forum, noone was able to help him).
Also the sorting order to find the latest bios on ECS website is idiotic.

P4S5A
Pro: With latest bios it works even with cpu 533 fsb and they added the 166 mhz frequency for both sdram and ddr, it allows to set very low timings without issues, I never had any single problem with it.
Con: Usb 1.1, ata-100, it doesn't support more than 1gb of ddr at 166 (memory test starts to throw errors as soon it starts to work with it), no official support for 533 cpus, unable to disable the parallel port on bios or windows xp will complain about not being acpi compliant, memory performance are lower than DX at same frequency and timings.

P4S5A/DX+
Pro: Usb 2, ata-133, official support to p4 3.06 ghz, better memory performance.
Con: more picky with memories, hyperthreading unusable, unable to disable the parallel port on bios or windows xp will complain about not being acpi compliant, unable to use 2gigs of ddr at 166mhz and tight timings.

Is there anyone with both mb that can confirm my issues?
Thanks.

yawetaG wrote:

[...]

My NLX system also has a SiS 5598 chipset (Asus SP98N motherboard), including the integrated video option, and although the board itself is generally stable the integrated video is anything but. It uses an early shared video memory implementation that is ...wonky..., to say the least. After changing the video memory size in the BIOS (it can be set from 1 Mb to 4 Mb) it takes several reboots before the memory change registers properly and the OS doesn't try to write data to that memory area. DirectX 5 is installed in Windows 98, but I daren't use it, because just redrawing the whole screen once when loading/exiting programs already shows massive slowdowns (you can see it redraw the screen 😵 ).

Good thing there are PCI slots. 😀

SiS pretty much invented shared memory video, the first PC UMA architecture was on the 5511 chipset, with the option to run a discrete 6202 VGA chip off its memory controller. Performance was of course abysmal. The 5596 basically integrated this into a single chip, then the 5598 added SDRAM support from the 5582 chipset to it. Still pretty primitive - although very variable between vendors, so I suspect BIOS implementation is as much to blame for things like you were seeing (although bad implementation guidelines are still SiS' fault).

SiS took a bit of a poison pill betting the company on UMA. They correctly saw that that was the future of mass-market computing, but by pushing it in the second half of the 1990s they ended up identifying themselves with bottom-scraping low-end solutions, with all the inherent quality and stability issues of that market segment. Eventually that killed them, despite the fact they frequently led the market in terms of innovation and performance - even when they had a truly excellent chipsets they couldn't charge enough for them to be profitable and they frequently only saw release on bottom-feeder boards. The SiS735 was probably their finest hour - the first time since the SiS 501 back in early 1994 that they could claim the absolute performance crown in the x86 market. Yet it was only available on a handful of arguably crappy boards, while even ALi's awful Magik1 chipset had better support.

SiS chipsets were a great choice in the early to middle 1990's. SiS 496/497 was one of the best chipsets for 486. Supported up to 1Mb L2 cache and the PCI bus. The last revisions made in 1996-1997 could actually work with EDO and there was a good chance to overclock them to 66MHz.

SiS lost most of their mainstream chipset market share to Intel and VIA in the late 1990's focusing on inexpensive solutions with integrated video based on their 6326 video chip. This silicon disaster didn't really deserve a favour of being called a GPU. They had to sacrifice even more performance in order to stuff everything into a single chip, so their Pentium compatible chipsets were the lowest end market. Their last chipsets of the early 2000's for Pentium 4 and Athlon were more competitive. They came up with a better SiS 315 GPU aka Mirage aka Real256 in 2001 which even supported T&L. The next move was to stuff this graphics into north bridge chips, so SiS 650 for Pentium 4 and SiS 740 for Athlon XP appeared. These were faster than Intel or VIA IGPs, though no match to the NVIDIA nForce2 IGP or ATI 9100IGP, but the poor drivers screwed them anyway.

dionb wrote:

yawetaG wrote:

[...]

My NLX system also has a SiS 5598 chipset (Asus SP98N motherboard), including the integrated video option, and although the board itself is generally stable the integrated video is anything but. It uses an early shared video memory implementation that is ...wonky..., to say the least. After changing the video memory size in the BIOS (it can be set from 1 Mb to 4 Mb) it takes several reboots before the memory change registers properly and the OS doesn't try to write data to that memory area. DirectX 5 is installed in Windows 98, but I daren't use it, because just redrawing the whole screen once when loading/exiting programs already shows massive slowdowns (you can see it redraw the screen 😵 ).

Good thing there are PCI slots. 😀

SiS pretty much invented shared memory video, the first PC UMA architecture was on the 5511 chipset, with the option to run a discrete 6202 VGA chip off its memory controller. Performance was of course abysmal. The 5596 basically integrated this into a single chip, then the 5598 added SDRAM support from the 5582 chipset to it. Still pretty primitive - although very variable between vendors, so I suspect BIOS implementation is as much to blame for things like you were seeing (although bad implementation guidelines are still SiS' fault).

SiS took a bit of a poison pill betting the company on UMA. They correctly saw that that was the future of mass-market computing, but by pushing it in the second half of the 1990s they ended up identifying themselves with bottom-scraping low-end solutions, with all the inherent quality and stability issues of that market segment. Eventually that killed them, despite the fact they frequently led the market in terms of innovation and performance - even when they had a truly excellent chipsets they couldn't charge enough for them to be profitable and they frequently only saw release on bottom-feeder boards. The SiS735 was probably their finest hour - the first time since the SiS 501 back in early 1994 that they could claim the absolute performance crown in the x86 market. Yet it was only available on a handful of arguably crappy boards, while even ALi's awful Magik1 chipset had better support.

I own an IBM Aptiva with the SiS 5511 chipset and the onboard SiS 6202 video, speaking of the SiS 735, it only went through few boards notably cheap brands, the only SiS 735 board i own is the ECS K7S5A

ph4nt0m wrote:

SiS chipsets were a great choice in the early to middle 1990's. SiS 496/497 was one of the best chipsets for 486. Supported up t […]
Show full quote

SiS chipsets were a great choice in the early to middle 1990's. SiS 496/497 was one of the best chipsets for 486. Supported up to 1Mb L2 cache and the PCI bus. The last revisions made in 1996-1997 could actually work with EDO and there was a good chance to overclock them to 66MHz.

SiS lost most of their mainstream chipset market share to Intel and VIA in the late 1990's focusing on inexpensive solutions with integrated video based on their 6326 video chip. This silicon disaster didn't really deserve a favour of being called a GPU. They had to sacrifice even more performance in order to stuff everything into a single chip, so their Pentium compatible chipsets were the lowest end market. Their last chipsets of the early 2000's for Pentium 4 and Athlon were more competitive. They came up with a better SiS 315 GPU aka Mirage aka Real256 in 2001 which even supported T&L. The next move was to stuff this graphics into north bridge chips, so SiS 650 for Pentium 4 and SiS 740 for Athlon XP appeared. These were faster than Intel or VIA IGPs, though no match to the NVIDIA nForce2 IGP or ATI 9100IGP, but the poor drivers screwed them anyway.

I have almost all chipsets except for the nForce2 and the Radeon 9000/9100 IGP, i also can say the IGP less SiS chipsets are quite decent, SiS 645 and 645DX are nice alternatives to the Intel 850/845 series and the SiS 655FX was even a rival to Intel 875P and 865 series

I also got the Netburst counterpart of the KM400 which is the VIA P4M800 (on an ASUS P4V8X-MX), i know also a better SiS IGP came on the SiS 760GX which adds DirectX 8.1 shaders support (although only pixel shaders), there no known SiS chipset for Intel CPUs having that IGP. i could bench and also pit it against the SiS 315 AGP card which i also have

Hello, some days ago I received a Radeon HD 4650 AGP, I wanted to try it on ECS P4S5A/DX+, everything went fine until I installed the windows xp driver, according to amd the only available drivers are: AMD_Catalyst_13.4_Legacy_Beta_WinXP, but when I tried them they were unable to find the card, so I used snappy driver installer, that installed a different version of drivers, the problem is that when I tried to reboot the system windows xp hanged up during the turn off phase, when I booted again it hangs during loading.
I thought the card was faulty, but it actually worked on a more updated rig (despite having troubles with driver once again).

Does anyone knows if there is any compatibility issue between such chipset and that cards?
Thanks.

probably not the sis chipset, possibly your driver is wrong and a borked install wrecked your system. possibly your card has problems.

"according to amd the only available drivers are: AMD_Catalyst_13.4_Legacy_Beta_WinXP" 🤣
Try to folow the instructions here. https://forums.tomshardware.com/threads/amd-1 … karound.718578/

Ah, I had the illustrious ECS K7S5A (SiS 735). I bought it because it supported SDRAM and DDR, and it just happened to be a great system. Fast and cheap! Those were the days!

On the other hand, that awful, awful 305 integrated graphics chip (on a different board). Or was it the 315? *vomit* It was all cool until you were stuck using that.

Install an AGP Hotfix version of Catalyst. v12.3 should work fine on XP.

my current build im working on

asus p4s533-e (sis 645dx)
pentium 4 2.53ghz 533fsb oc @ 2.907ghz 153fsb 1.6vcore
1 stick corsair xms 3200 512mb 2.5-3-3-7 timings mem freq @ 333 (400 is unstable when using 533fsb p4 vs 400fsb p4)
voodoo 5 agp
turtle beach santa cruz
80gb ide seagate

pros
stable board simple layout
fast under 98se duh
beats 845e competitive against 850/e rdram 800 setups
arguably the better setup for a fast v5 compared to the played out kt333 epox
p4 3.067ghz works and hits 3.4ghz even more overkill
universal agp support for 3.3v cards
vore upto 1.8v
vdimm upto 2.9v
mem freq of 200,266,333 and unofficial 400mhz

cons
no agp/pci lock
153 fsb wall chipset limitation bios goes upto 166fsb
not much else negative to say

still testing the board but im happy so far. was gonna go period correct but the p4 really opens it up for the v5 so im good im already spoiled. plan on buying another 645dx for my kyro2 build as well.

By any chance you have been able to enable the HT on asus p4s533-e?

Nemo1985 wrote:

By any chance you have been able to enable the HT on asus p4s533-e?

no sir if im correct sis didnt give there p4 boards ht support till 651/fx i think. i know im probably missing some other matx board that probably have ht support but eh. 533 p4 with ht is better on a granite bay or 850e in my opinion.

*edit also sis 655 is another chipset and probably there best as far as p4 boards. came out late borderline 02 early 2003. eh but by then springdale showed its dominance.

greasemonkey90s wrote:

Nemo1985 wrote:

By any chance you have been able to enable the HT on asus p4s533-e?

no sir if im correct sis didnt give there p4 boards ht support till 651/fx i think. i know im probably missing some other matx board that probably have ht support but eh. 533 p4 with ht is better on a granite bay or 850e in my opinion.

*edit also sis 655 is another chipset and probably there best as far as p4 boards. came out late borderline 02 early 2003. eh but by then springdale showed its dominance.

Thank you for the confirmation, my sis mb has the HT option in the bios but it's heavily bugged, when turned on the screen is full of glitches.

Nemo1985 wrote:

greasemonkey90s wrote:

Nemo1985 wrote:

By any chance you have been able to enable the HT on asus p4s533-e?

no sir if im correct sis didnt give there p4 boards ht support till 651/fx i think. i know im probably missing some other matx board that probably have ht support but eh. 533 p4 with ht is better on a granite bay or 850e in my opinion.

*edit also sis 655 is another chipset and probably there best as far as p4 boards. came out late borderline 02 early 2003. eh but by then springdale showed its dominance.

Thank you for the confirmation, my sis mb has the HT option in the bios but it's heavily bugged, when turned on the screen is full of glitches.

np if im correct the abit sa7 which looks to be a relabled version of the ecs board you have does support the 3.067ht with ht enabled. it was confirmed on a voodooalert thread i googled before i bought this p4s533-e. no idea why its buggy on the ecs p4s5a/dx+.

i have an Elitegroup ECS 755-A2 1.0A with SIS chipset and it runs fine without problems. It has a Sempron 1,6GHz and 512Mb of ram. Its my only purple board, I like different colored boards not only those boring brown and green.

My 486 VLB motherboard (Asus VL/I-486SVGO VLB board) is extremely fast and compatible, based around a SIS chipset.

When I had a K75SA based motherboard, based on the SIS 735 chipset, that was really good as well.

I also have a P4 SiS based motherboard, A gigabyte GA-8S648FX, based on the SiS 648FX. Not as fast as a dual channel Intel 845 board, but better DOS compatibility.

I have an experience with Sis 735 - K7S5a, one of the best boards from early SocketA era. Also with K7S6A, also very good.

But best for me was P6s5at - socket 370 board with DDR support. I had Abit BE6 with Celeron 750 in slot1 adapter. Overclocking was very limited about 850MHz was its limit with all the overclocking capabilities on Abit board (probably also because of slot1 adapter). In sis board P6s5at celeron run stable at 1150 MHz without any problems and without even voltage adjustment (that board had no such setting)

I had a SiS mAtx motherboard that was stable. But even though it had universal Agp, some agp revisions didn't work. I also flashed the latest bios to no avail.
It didn't worked with my V3 after installing the driver of the video card. Just blank screen after the reboot. I prefer Via Universal Agp boards. More stable and way more features than Sis in the Bios.

So, what is latest SIS chipset, that supports Universal AGP (3.3v signaling), as far as I see, SIS 645dx is ok, but doesnt support 800 Mhz FSB and HT, and Prescott CPUs.
651 seems same as 645dx, only better integrated graphics? Also built in support for USB 2.0. What about 648fx? This supports HT for sure, 800 Mhz p4 too, but not sure if there are some boards, that supports Universtal AGP 2x,4x, so 3.3 v signaling. Because Asus P4S800 doesnt support it, only 1.5v AGP cards. Not sure, if this apply to all 648fx motherboards.
Anyway, what's last SIS chipset for P4, that can support AGP 2x 3.3v cards?

I had SiS620-based Socket-370 mainboard by Jetway in late 90s. Typical cheap mainboard with poor 3D capabilities simply integrated graphics.