VOGONS


First post, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I have an AZZA PT-5IT socket 7 motherboard (not a super7). It has 512 KB pipeline burst write-back cache that can accept 256 MB SDRAM (DIMM) or 256 MB EDO RAM (SIMM). Of course, it can only cache up to 64 MB as there is no write-through option in the BIOS.

It uses the Intel 430TX chipset and has options for 50, 60, 66, 75 Mhz FSB and 1.5, 2.0, 2.5, 3.0, 3.5, and 4.0 multipliers. It supports a wide range of voltages, 2.1, 2.5, 2.8, 2.9, 3.2, 3.3, 3.5 V.

The two questions I have are,

1) Are there hidden FSB settings besides what I postd above? There are 4 other possible combinations for jumpers. An ~83 Mhz setting would be ideal.

2) The motherboard seems to work with both the K6-266 @ 266 Mhz and Cyrix MII-400GP @ 263 Mhz. Which of these two cpus is considered the best all around performer? The reason for questioni 1) is to try to get the Cyrix MII-400GP to run at its rated frequency of 285 Mhz. I attempted the 75 Mhz w/4x but the BIOS reported 66 Mhz and didn't POST.

3) Are there any architectural differences between the Cyrix MII-400GP and the Cyrix 6x86MX-266/or MII-266 that the BIOS might need to know about for optimal performance?

Reply 1 of 17, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I think I have found the answer (I ran some tests).

Cyrix MII-400GP @ 262.5 Mhz

Quake Timedemo demo1 (fps): 42.4
Doom Timedemo demo 3 (fps): 103
3DBench 1.0c: 363
Pcpbench: 30.2
Cachechk v4 (L1,L2,RAM (MB/s)): 1092, 227, 118
DHRY1OD (VAX MIPS): 402
LINPCOD (MFLOPS): 28
WHETCOD (MFLOPS): 123
BYTEmark32 (P90 ratio, Int/FP): 3.51, 1.95
Jbench (ALU/FPU): 31.7, 31.4
Landmark (ALU/FPU): 3133, 3573
PiDOS (25k digits, s): 6
CPUIndex: 43
PC-Config v9.33: 333
Sysinfo: 5055

AMD K6-2-266 @266 Mhz

Quake Timedemo demo1 (fps): 42.4
Doom Timedemo demo 3 (fps): 89.8
3DBench 1.0c: 234.6
Pcpbench: 27.5
Cachechk v4 (L1,L2,RAM (MB/s)): 1078, 278, 106
DHRY1OD (VAX MIPS): 451
LINPCOD (MFLOPS): 27.4
WHETCOD (MFLOPS): 166.1
BYTEmark32 (P90 ratio, Int/FP): 3.84, 2.46
Jbench (ALU/FPU): 26.7, 26.1
Landmark (ALU/FPU): 2731, 4318
PiDOS (25k digits, s): 4
CPUIndex: 48
PC-Config v9.33: 333
Sysinfo: 1582

From the above data, it appears as if the AMD is very marginally better, as an average, than the Cyrix at ALU performance and the Cyrix is slightly better at FPU performance.

I found the 83 Mhz FSB motherboard setting and was able to run the Cyrix MII-400GP at 292 Mhz. For older Socket7 motherboards, this may be the best overall option available (as well as the MII433GP @ 292 Mhz).

Quake Timedemo demo1 (fps): 47.2
Doom Timedemo demo 3 (fps): 113.5
3DBench 1.0c: 404
Pcpbench: 33.6
Cachechk v4 (L1,L2,RAM (MB/s)): 1216, 253, 132
DHRY1OD (VAX MIPS): 448
LINPCOD (MFLOPS): 31
WHETCOD (MFLOPS): 137
BYTEmark32 (P90 ratio, Int/FP): 3.91, 2.16
Jbench (ALU/FPU): 35.2, 35
Landmark (ALU/FPU): 3488, 3974
PiDOS (25k digits, s): 5
CPUIndex: 48
PC-Config v9.33: 370

Reply 2 of 17, by Tetrium

User metadata
Rank l33t++
Rank
l33t++

Looking at the voltage options, I'd say there could be more undocumented voltage settings available (depends on how the other voltages are jumpered.

Whats missing in your collections?
My retro rigs (old topic)
Interesting Vogons threads (links to Vogonswiki)
Report spammers here!

Reply 3 of 17, by Tetrium

User metadata
Rank l33t++
Rank
l33t++

Heres what I mean, the empty spots may have another usable voltage to pick from 😉

Attachments

Whats missing in your collections?
My retro rigs (old topic)
Interesting Vogons threads (links to Vogonswiki)
Report spammers here!

Reply 4 of 17, by feipoa

User metadata
Rank l33t++
Rank
l33t++

There may be more voltage options, but I'm not sure if any other voltage is of interest for early-to-mid date socket 7 cpus. It is fairly straight forward to modify the voltage regulators anyway. What other voltages are of interest for socket 7 cpus below 300 Mhz that are not in that list?

The pairing of jumpers for different voltages may makesuse of parallel and serialising of resistors to the multiple on-board voltage regulators. I'm not sure why AZZA put on 3 different regulators; multiplexing would be easy enough with 1 regulators (maybe for current summation?). If not as a current adder, would they be outputting different voltages from different regulators? That doesn't make sense to me. Hmmm, maybe there are other items on the board that need precise voltages. I supposed I could trace out the circuit, but I'm not so interested in this 'fast' Pentium stuff.

I was a little disappointed that the last of the real Cyrix cpus (MII-433GP) would probably only equate to that of an AMD K6-2-300. At least it would be on-par clock-for-clock. There may be a few MII-466GP's out there which may be even more evasive than the elusive Cyrix 5x86-80/133's. I'd like to throw the Cyrix 5x86-120/4X into that category as well. I'll stop ranting about rare Cyrixes now!

EDIT: I'll answer my own question again -- I can see two regulators for MMX cpus, but not a need for 3, at least not for the cpu. Ok, I got it... to drop the RAM voltage to 3.3 V when a 3.5V core cpu is inserted.

Reply 5 of 17, by feipoa

User metadata
Rank l33t++
Rank
l33t++

Found a few other settings, the tally on this 430TX is now:

FSB (Mhz)
50
55
60
66
75
83

Multiplier (x)
1.5
2.0
2.5
3.0
3.5
4.0
4.5

So using PC100 SDRAM over PC66 or SIMMs, it may be possible to run a cpu, without circuit mods to the board, at 83.3 x 4.5 = 375 Mhz, or an AMD K6-380. This is pretty good for an intel-based non-super socket7 without an Evergreen or Kingston upgrade kit.

I've read online that a BIOS update would make the K6-III usable logic wise and multiplier wise. Apparently, a 2.0x gets transferred to a 6.0x setting. So 75 x 6.0 = K6-III-450, or to OC the K6-III, 83.3 x 6.0 = 500 Mhz.

Does anyone have, or know where to find, a BIOS update for the AZZA PT-5IT? Thanks!

EDIT: Apparently Azza never made a BIOS update, so it looks like a K6-2 is the latest available. The frequency generator circuit may be replaceable with a chip that will go up to 100 Mhz and probably a 5.5x multiplier, allowing for the K6-2-550.

Reply 6 of 17, by unmei220

User metadata
Rank Member
Rank
Member

Bear in mind that the Intel 430TX chipset doesn't support Linear Burst mode which, when enabled, supposedly gives the Cyrix CPUs a 5-10% boost.

In order to get the best out of the 5x86, the system logic should support write-back caching, burst writes and the Cyrix Linear Burst sequence. Intel patented the burst write order of their CPUs, and only AMD had the rights to copy that patented burst order. Cyrix made their own, called Linear Burst otherwise they would have had to spend additional clock cycles with the 1+4 Intel-compatible burst order.

I have two TX mobos, a cheap PCChips M550 and a Soyo 5BT. The MII refuses to work correcly on the Soyo, but works perfectly on the M550. For the M550 to do 75Mhz FSB stable, I had to add a small heatsink to the northbridge.
Maybe you can send your BIOS to Jan Steunebrink, if he has time maybe he can add K6-2+/III+ support for your board. I contacted him recently about the M550 mobo, which now sports full K6-2+/III+ support.
http://web.inter.nl.net/hcc/J.Steunebrink/k6plus.htm

Reply 7 of 17, by feipoa

User metadata
Rank l33t++
Rank
l33t++

Thanks for the link, I'll give Jan a call.

How did you determine that linear burst wasn't working on the MII cpus? Is this a factor of all i430TX chipsets? Do some BIOSes have a workaround for this to enable LINBRST functionality?

The oddity I've noticed with MII's and cyrix MX cpus is with CTCM, it says, "results for Cx686 with Cache Fill on Write miss not correct!" Quake and Doom seem to run fine though. Speedsys 4.78 won't load though, it hangs on MSR something.

Reply 8 of 17, by unmei220

User metadata
Rank Member
Rank
Member

Intel chipsets just don't support Linear Burst. Chipsets from other manufacturers like VIA, SiS, OPTi, ALi, etc almost all support it. I couldn't find a tool or a way to know if this feature is enabled by default or you have to make something to enable it. For example, I have another PCChips M571 board which uses SiS 5598 chipset. That chipset supports Linear Burst, but I don't know a way to check if it's enabled or not. I made a thread here recently about it: Cyrix MII CPU - Linear burst mode

For a complete list of chipsets that support Linear Burst, check this page:
http://mysite.verizon.net/pchardwarelinks/chi … ets_pentium.htm
(look for the "Lin Burst" string)

Reply 10 of 17, by meljor

User metadata
Rank Oldbie
Rank
Oldbie

this is a 2011 topic...

asus tx97-e, 233mmx, voodoo1, s3 virge ,sb16
asus p5a, k6-3+ @ 550mhz, voodoo2 12mb sli, gf2 gts, awe32
asus p3b-f, p3-700, voodoo3 3500TV agp, awe64
asus tusl2-c, p3-S 1,4ghz, voodoo5 5500, live!
asus a7n8x DL, barton cpu, 6800ultra, Voodoo3 pci, audigy1

Reply 11 of 17, by feipoa

User metadata
Rank l33t++
Rank
l33t++
sliderider wrote:

Is this meant to be a test of chips running strictly at or below 300mhz actual speed? Have you tried running the 400GP at 100x3?

Yes, refer to 686 benchmark comparison.

Plan your life wisely, you'll be dead before you know it.

Reply 13 of 17, by meljor

User metadata
Rank Oldbie
Rank
Oldbie

No need to get offensive, i'm fairly New here.
On other forums it isn't that normal you'll get an answer after a couple of years.

My bad...

asus tx97-e, 233mmx, voodoo1, s3 virge ,sb16
asus p5a, k6-3+ @ 550mhz, voodoo2 12mb sli, gf2 gts, awe32
asus p3b-f, p3-700, voodoo3 3500TV agp, awe64
asus tusl2-c, p3-S 1,4ghz, voodoo5 5500, live!
asus a7n8x DL, barton cpu, 6800ultra, Voodoo3 pci, audigy1

Reply 14 of 17, by carlostex

User metadata
Rank l33t
Rank
l33t
meljor wrote:

No need to get offensive, i'm fairly New here.
On other forums it isn't that normal you'll get an answer after a couple of years.

My bad...

There's really no need to apologise. Some people are just selective about which threads they like to get revived or not. Personally i don't give a damn.

Reply 15 of 17, by feipoa

User metadata
Rank l33t++
Rank
l33t++

It is my personal preference that old threads be revived rather than starting new threads of the same topic, irregardless of how old the thread creation date is.

Plan your life wisely, you'll be dead before you know it.

Reply 16 of 17, by carlostex

User metadata
Rank l33t
Rank
l33t
feipoa wrote:

It is my personal preference that old threads be revived rather than starting new threads of the same topic, irregardless of how old the thread creation date is.

Precisely my point, but sometimes some might have double standards.

Reply 17 of 17, by Stiletto

User metadata
Rank l33t++
Rank
l33t++
carlostex wrote:
feipoa wrote:

It is my personal preference that old threads be revived rather than starting new threads of the same topic, irregardless of how old the thread creation date is.

Precisely my point, but sometimes some might have double standards.

I guess I'm considered "having double standards"? Here's my point of view...
Re: Tex's new Socket 7 DOS machine

"I see a little silhouette-o of a man, Scaramouche, Scaramouche, will you
do the Fandango!" - Queen

Stiletto