VOGONS


First post, by PlaneVuki

User metadata
Rank Member
Rank
Member

Hi.

I have this HMC isa vga card. Initially I thought that it is simply a slow card, like those oak ones. But then it seemed really slow. So I had a suspicion and tried it in 8bit mode, with 16bit pins covered with paper. And it yielded same slow result in doom benchmark. So I think it works in 8 bit mode only whether used in 8bit or 16bit isa slot.

How can I make it work in 16 bit mode?
Anyone has manual for this video card?

It seems to have 3 jumpers, jp1 (0ws), jp3 (int) and jp5.
Jp1 and jp3 made no differece in doom.
Jp5 is a multipin single jumper, despite what looks like to be many jumpers. The jumper is one big part, so all either shorted at 12 or all shorted at 23. Shorting it at 23 results in kinda low color mode in doom. Like 16 colors or something. No difference in performance.

What can be done?

Thanks a lot.

Attachments

Reply 1 of 17, by wierd_w

User metadata
Rank Oldbie
Rank
Oldbie

This appears to be a "newer" version of the card listed in TULARC.

https://stason.org/TULARC/pc/graphics-cards/U … 4-VER-1-01.html

There is a jumper called '0 wait-state', (JP1 in your image, JP24 in the TULARC doc) that might be worth investigating.

You have the 40/44mhz clock jumper pads present, but no header is installed on the pads.
(You would likely need a 44mhz crystal subbed for that 40mhz one if you did that anyway...)

The other jumpers seem to define the DRAM configuration, and the use of IRQ9.

Reply 2 of 17, by MikeSG

User metadata
Rank Member
Rank
Member

It looks like the main processor is an 8-bit chip judging by it's size. If placing a jumper on 0-Wait-State does nothing, you could try using Bus Clock Divider settings in Bios if present. Eg. BusClk / 2.

Reply 3 of 17, by bakemono

User metadata
Rank Oldbie
Rank
Oldbie

Doesn't DOOM use only 8-bit writes in the first place? I think it would be better to use a real video memory benchmark to test this behavior.

again another retro game on itch: https://90soft90.itch.io/shmup-salad

Reply 4 of 17, by PlaneVuki

User metadata
Rank Member
Rank
Member
bakemono wrote on 2024-04-07, 12:26:

Doesn't DOOM use only 8-bit writes in the first place? I think it would be better to use a real video memory benchmark to test this behavior.

I am very sure that it uses 16bit writes. I have tested doom with many other cards, both in 16bit and 8bit isa modes, 8bit mode speed is considerably slower than 16bit mode. The difference is half when using fast cpu since isa vga is huge bottleneck in such case.

Reply 5 of 17, by konc

User metadata
Rank l33t
Rank
l33t

There are some cards that are compatible with an 8-bit slot and are actually 8-bit cards, MikeSG has a point.

I'm not saying that this particular card is one of those, I don't know. I'm only saying that this is a possibility to consider.

Reply 6 of 17, by waterbeesje

User metadata
Rank Oldbie
Rank
Oldbie

Some variants of the notoriously slow Trident 8900B have the same issue. There may be a jumper for 8/16b operation, but the 16b setting actually triggers an auto detect system that isn't very great at it. So it always remains in 8b mode.

Looking at your card, the 100ns ram does not help a lot as well... But could be worse.

Stuck at 10MHz...

Reply 7 of 17, by rasz_pl

User metadata
Rank l33t
Rank
l33t
PlaneVuki wrote on 2024-04-07, 16:16:

I am very sure that it uses 16bit writes.

Afaik original doom indeed does half of the writes (columns=walls) in 8bit mode, it draws them one pixel line at time.

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 8 of 17, by DEAT

User metadata
Rank Newbie
Rank
Newbie

I have a HMC HM81314Q card. It is definitely a 8-bit only card, as none of the 16-bit pins have any traces leading to them. Like Realtek RTG3105iEh cards, it was designed more around its form factor rather being practical.

In Windows 3.1 benchmarks, it's literally the slowest card that runs at 800x600x256 - even the Realtek RTG3105iEh and Trident 9000i are faster.

Reply 9 of 17, by PlaneVuki

User metadata
Rank Member
Rank
Member

Well there clearly is a performance hit in 8bit mode. For example I had this slower et4000 isa vga in past, looking at results when used on socket 7 and pentium 200 (isa vga is bottleneck):

8bit mode: 12.8 fps
16bit mode: 22.4 fps

Anyway. So nothing can be done for this hmc card then right. It is 8bit only it seems.

Reply 10 of 17, by wierd_w

User metadata
Rank Oldbie
Rank
Oldbie

Yes.

8bit mode (assuming an XT), would get penalized both by the datapath being half that of the 16bit mode, AND from the bus speed being ~25% slower than the 6mhz rate of the 6mhz 16bit AT bus.

However, this requires comparing apples to oranges--
If you are comparing a benchmark that is straight 8bit writes for both, you end up just measuring the increase in speed of the bus.

If you are comparing 8bit writes to 16bit writes, this is apples to oranges; you are measuring the increase in speed of the software, by leveraging a larger word size on the databus. Assuming 0 waitstates, and ideal bus optimization, you can expect a 100% increase in thruput from writing 16bit words instead of 8bit words, per clock cycle. (It is further complicated in the real world, by some CPU operations requiring more than one clocking cycle to complete. For the same instruction being called, the actual real-world difference in time to complete it may be quite drastically different, between a CPU that has implementations of that instruction that complete in 2 cycles instead of 3, or 1 cycle instead of 2, for example. Some of the instructions that got changed between 286 and 386, for instance, had such improvements in cycles to complete, in addition to the wordsize and databus width being made larger. Similar story for 386 to 486, and 486 to pentium. Some versions of some chips had differences that made big impacts in this area, like the 386slc vs 386sx, which had 16bit external bus vs 32bit external bus, and thus would have waitstates or multiple clock cycles needed to do a 32bit write, in the case of the narrower bus variant. There's a lot of devils in the details with some arbitrary benchmarking methods.)

If you are comparing 8bit writes on an XT's 4.77mhz bus, against 16bit writes on a 386's 8mhz bus, you are very much comparing very different beasts straight up. All things being considered ideal, you are looking at ~360% increase in theoretical performance. (Possibly larger, when you get things like cache involved, but that would mostly only happen on READS, not writes.)

[NOTE-- this is meant as "Comparison of rate you can redraw the screen between an actual XT, and a 386 AT clone."]

As for having a slow ISA video card-- remember, these devices had ... Conflicting... intentions in the marketplace with what we want them for now. Lots of commodity parts were made for the purposes of displaying boring business data on the screen, and little else. How fast you can redraw the screen is immaterial for such applications. WordPro, WordPerfect, Lotus 1-2-3, and pals really just wanted the higher display modes to cram more text / data on the screen at once. Not to get greater than 30fps in a raycasted FPS game. 😜

The PC very very much wanted to be seen as a business instrument, and not a glorified game console, you see.

Reply 11 of 17, by rasz_pl

User metadata
Rank l33t
Rank
l33t
wierd_w wrote on 2024-04-08, 08:01:

8bit mode (assuming an XT), would get penalized both by the datapath being half that of the 16bit mode, AND from the bus speed being ~25% slower than the 6mhz rate of the 6mhz 16bit AT bus.
Assuming 0 waitstates

its even worse, XT 8 bit ISA had no NOWS pin on the slot so no 0 waitstates
fastest 8bit AT ISA transfer is 3 cycles, fastest 16bit AT ISA transfer is 2 cycles

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 12 of 17, by PlaneVuki

User metadata
Rank Member
Rank
Member
rasz_pl wrote on 2024-04-08, 13:06:
wierd_w wrote on 2024-04-08, 08:01:

8bit mode (assuming an XT), would get penalized both by the datapath being half that of the 16bit mode, AND from the bus speed being ~25% slower than the 6mhz rate of the 6mhz 16bit AT bus.
Assuming 0 waitstates

its even worse, XT 8 bit ISA had no NOWS pin on the slot so no 0 waitstates
fastest 8bit AT ISA transfer is 3 cycles, fastest 16bit AT ISA transfer is 2 cycles

So since my hmc card has 0ws jumper, it should be a 16bit card but somehow works only in 8bit mode? 0ws jumper has absolutely no effect on performance in doom benchmark. I tested many times.
I want it to work in 16bit mode for a more decent frame rate. Now it is very bad, around 11-12fps when used in either socket7 or socket3 systems.

Reply 13 of 17, by rasz_pl

User metadata
Rank l33t
Rank
l33t

Its not about the card, but motherboards. XT computers (8 bit slot) didnt have 'no waitstates' signal defined, AT boards (16 bit slot) repurposed one of the pins.

Open Source AT&T Globalyst/NCR/FIC 486-GAC-2 proprietary Cache Module reproduction

Reply 14 of 17, by Jo22

User metadata
Rank l33t++
Rank
l33t++
rasz_pl wrote on 2024-04-08, 15:24:

Its not about the card, but motherboards. XT computers (8 bit slot) didnt have 'no waitstates' signal defined, AT boards (16 bit slot) repurposed one of the pins.

It seems to be pin B 08, which had been reserved on PC and PC/XT.

ISA bus:
https://de.wikipedia.org/wiki/Datei:ISA_Bus_pins.svg

PC bus:
https://de.wikipedia.org/wiki/Datei:XT_Bus_pins.png

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 15 of 17, by DEAT

User metadata
Rank Newbie
Rank
Newbie
PlaneVuki wrote on 2024-04-08, 13:40:

I want it to work in 16bit mode for a more decent frame rate. Now it is very bad, around 11-12fps when used in either socket7 or socket3 systems.

Take a closer look at the photos you posted of your card - there is literally no traces going to the 16-bit pins, and I imagine it will be the same on the other side of the card. This is exactly the same as my HM86314Q card.

You're not going to get 16-bit ISA speeds on that card, period.

Reply 16 of 17, by PlaneVuki

User metadata
Rank Member
Rank
Member
DEAT wrote on 2024-04-10, 22:53:
PlaneVuki wrote on 2024-04-08, 13:40:

I want it to work in 16bit mode for a more decent frame rate. Now it is very bad, around 11-12fps when used in either socket7 or socket3 systems.

Take a closer look at the photos you posted of your card - there is literally no traces going to the 16-bit pins, and I imagine it will be the same on the other side of the card. This is exactly the same as my HM86314Q card.

You're not going to get 16-bit ISA speeds on that card, period.

Yes I actually looked closer for traces and found none. But I thought maybe the traces are beyween the layers of pcb.

Thanks. I accept the fact that it is 8bit only. But why fake the rest I dont understand.