feipoa wrote:

I'm not sure what those extra DIP holes are for next to the DIP-32 sockets. Board photo, http://img189.imageshack.us/img189/1420/oxqi.jpg

TAG ram socket?

I am not sure. I have never seen a TAG RAM socket spaced out over 4 DIPs like this. I suppose it is possible, but it wouldn't be a 32kx8 like the manual mentions.

feipoa wrote:

I am not sure. I have never seen a TAG RAM socket spaced out over 4 DIPs like this. I suppose it is possible, but it wouldn't be a 32kx8 like the manual mentions.

It's DIP-28 cache chips placed in DIP-32 sockets. It's common to see that on Socket 5/7 boards, might not be on 486s

To clarify what I am refering to, view attached image.

Ahh.. I think that was referenced in the manual. Something about using the wider SRAM DIPS for the larger capacities.

EDIT: Page 4 - 9 "Bank 0 is made up of dual sockets that can accept the larger cache chips used in the 512KB cache"

I originall thought that the use of "larger" in that sentance was refering to DIP-32 vs. DIP-28 cache chips, which should not have any bearing on these 4 extra 16-pin row of sockets. In case you are not familiar, DIP-32 means 32 pins of the DIP is the style (long pointy pins).

Ahh, ok so in this case dual = longer, not wider. I wasn't sure after reading it.

From my understanding, "dual" does not equate to "longer". I'm not sure what they mean by "dual sockets", but I suspect it has something to do with those 4 extra rows of 16-pin DIP sockets. It seems like user Anonymous Coward had some board in the past with these, but I personally do not know how to fill them.

Artex wrote:

vetz wrote:

I bought them from aliexpress.com. It was cheaper than Ebay. I bought from this seller: http://www.aliexpress.com/item/IS61C1024-10N- … 1832174608.html. They were working chips, but they are advertised as 10NS, I received 15NS. Makes little difference in practice, but try another seller (or get partial refund).

Has anyone sourced these chips from other places? This is gonna be expensive for all these chips.

I took a break from that pain-in-the-ass SP3G board and threw the AMD X5-133 into my second SV2GX4 board (Rev 2.0). I basically modified build #2, swapping the Intel DX4-100 Write-Back with this processor. L2 Cache is set to AUTO in the BIOS and L1 cache is set to WB mode with a jumper.

Build 2:
Asus VL/I-486SV2GX4 Rev. 2.0 running latest available BIOS (0402.001 BETA BIOS from 1999)
AMD-X5-133ADZ (Am5x86-P75) - 133Mhz - 33Mhz - 16KB WB (11/6/1995)
16MB (Was 32MB RAM)
256KB L2 Cache
TSENG ET4000/W32P - Diamond Stealth 32 2MB VLB (1993)
4GB CF Card (CF->IDE Adapter) on VLB Promise Controller
GOTEK USB Floppy Emulator

System 2 Benchmark Results:
SpeedSys Overall Score: 50.25
DOOM: 2134 gametics in 1425 realtics
Superscape: 80.1 fps
PC Player Bench: 20.3
Quake: 969 Frames 76.1 seconds 12.7 FPS

Cache Performance:

CacheChk -d -t6
L1 (16KB) - 137.3 MB/s
L2 (256KB) - 62.0 MB/s
Main Memory Speed - 42.4 MB/s
Effective RAM Access Time (read) - 98ns
Effective RAM Access Time (write) - 61ns

CacheChk -d -w -t6
Main Memory Speed - 68.8 MB/s (15.2ns)
Effective RAM Access Time (write) - 60ns

Artex wrote:

I took a break from that pain-in-the-ass SP3G board and threw the AMD X5-133 into my second SV2GX4 board (Rev 2.0). System 2 Be […]
Show full quote

I took a break from that pain-in-the-ass SP3G board and threw the AMD X5-133 into my second SV2GX4 board (Rev 2.0).
System 2 Benchmark Results:
SpeedSys Overall Score: 50.25
DOOM: 2134 gametics in 1425 realtics
Superscape: 80.1 fps
PC Player Bench: 20.3
Quake: 969 Frames 76.1 seconds 12.7 FPS

Your results just confirm that the SV2GX4 is a kickass board when it comes to performance. You just beat all the PCI (and VLB) boards benched with the same CPU with those benches 😀

Just imagine with 1024kb L2 cache 😲

Interesting results for sure. Do the cache/memory speeds look ok? Starting to change my mind on these builds given my experience with these boards..

Build 1: 1 x BIOSTAR MB-8433UUD-A Rev. 3.1 with AMD-X5-133ADZ (Am5x86-P75) in Write-Back Mode
Update 06/13 - Might try running the ADZ @ 160Mhz - benchmarks coming....

Build 2: 1 x Asus VL-i486SV2GX4 Rev. 2.0 with AMD-X5-133ADZ in Write-Back Mode
Update 06/13 - Swapped the Intel DX4-100 for this one and got some really good results (see above). Might upgrade cache to 1MB.

Build 3: 1 x Asus VL-i486SV2GX4 Rev. 2.1 with Cyrix 5x86 120 in Write-Back Mode
Update 06/13 - Need to figure out reason for slow memory/cache performance???

Build 4: 1 x Asus PCI I-486SP3G Rev. 1.8 with Intel i486DX4-100 &EW in Write-Back Mode
Update 06/13 - Swapped the AMD X5-133ADZ for the Intel i486DX4-100 &EW. Need to verify WB mode works and bench.

Yes, those memory results are as I would expect. They are the same impressive results you get on a SiS-based PCI 486.

Are you able to test the two undocumented FSB settings on that SP3G? If it is 40 MHz, that would make the board usable.

Well gents... I feel like a jackass. I switched gears again to benchmark the BIOSTAR board and after getting everything up and running at 160Mhz (stable), I decided to flash the EEPROM... which failed...

BECAUSE IT WAS AN EPROM (UV-erasable), non-EEPROM. 🙁 I should have checked first (facepalm). 😵

So... now I assume I need to A) Get a programmer and B) get an EEPROM?

I guess this gives me more time to try out the undocumented SP3G jumper configs?

Exactly! The SP3G board, in case it was not profoundly obvious, has caught my attention. Check for 40 MHz.

I did not realise a 3.1 version of the Biostar board came with a UV EPROM. I've read on here that people leave it outside on a sunny summer day and they can erase it. Maybe you could try this and then reprogram it?

Do you really need to buy a programmer? Could you not just buy the EEPROM and hot flash it in another motherboard? That is what I did before I acquired a programmer.

feipoa wrote:

Do you really need to buy a programmer? Could you not just buy the EEPROM and hot flash it in another motherboard? That is what I did before I acquired a programmer.

How does that work? You buy the EEPROM, throw it in another board, boot it up with a floppy that has the flashing utility & BIOS and flash it?

EDIT: NVM, didn't realize the "HOT" part meant that the chips are swapped while the system is ON. 😀 I'll need to grab a chip removal tool and one (or a few) of these in case something goes wrong.

feipoa wrote:

The SP3G board, in case it was not profoundly obvious, has caught my attention. Check for 40 MHz.

I had a SP3G once, there was no 40 MHz setting, otherwise I'd probably kept it...
Remember this has an Intel chipset, so the incentive to support >33 MHz FSB was rather low, since pretty much only CPUs from Intels competitors would support higher FSBs.

Alrighty.. so I found another board that has a Winbond W29C011A-15 chip. I was able to flash it using the latest Biostar 2012 BIOS, and I dropped this chip into the MB-8433UUD board. I powered it on and.. success...or so I thought. Everything is REAL sluggish - and the benchmarks are coming back showing that as well. I mean... real slow.

Did I "eff" something up here?

5u3 wrote:

feipoa wrote:

The SP3G board, in case it was not profoundly obvious, has caught my attention. Check for 40 MHz.

I had a SP3G once, there was no 40 MHz setting, otherwise I'd probably kept it...
Remember this has an Intel chipset, so the incentive to support >33 MHz FSB was rather low, since pretty much only CPUs from Intels competitors would support higher FSBs.

Do you recall what FSB speeds corresponded to the other two undocumented jumper settings?

While the Intel chipset may not officually support 40 MHz+, the motherboard manufacturer, Asus, may still have sourced a clock generation chip which is configured for many of the standard frequencies at the time - 25, 33, 40, and 50 MHz. Similar to how UMC boards contain clock generation chips which have 25, 33, 40, 50, 60, 66, 83 MHz options.

Artex wrote:

Alrighty.. so I found another board that has a Winbond W29C011A-15 chip. I was able to flash it using the latest Biostar 2012 BIOS, and I dropped this chip into the MB-8433UUD board. I powered it on and.. success...or so I thought. Everything is REAL sluggish - and the benchmarks are coming back showing that as well. I mean... real slow.

Did I "eff" something up here?

Be sure to set the appropriate jumper for EEPROM 5V, not 12V. There are jumper settings for this. Please check my manual for this board. Slow results may be because you have the Dirty TAG bit set incorrectly. Please use 7+1 when the L2 cache is in write-back mode and 8+0 when the L2 cache is in write-thru mode.

EDIT: After doing this, it is a good idea to set all BIOS settings to default. If that fails, then try to flash the 2012 BIOS while in the board with uniflash.