VOGONS


Reply 20 of 200, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Are there any BX boards with UDMA66 and 3 ISA (or more) ISA slots?

No.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 21 of 200, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
Doornkaat wrote on 2022-02-16, 15:24:

Can do, will do!👍 (Please remind me if I forget.😅)
Since the program seems to do much more than just controlling the PLL I wasn't sure if you also wanted to know anything specific to the VIA C3 on this board.
I'll see if I have a CPU here that's likely to run at 133MHz FSB. The 1.1GHz will probably crash immediately.😄

Awesome! 😁
Yeah, you'll need a CPU that's either natively running on 133 MHz, or one that can handle the overclock (I have some 700 MHz Pentium 3 Coppermine CPUs that run just fine at 933 MHz, without additional voltage).
Or, you can just test other frequencies, that PLL handles quite a few: 66, 75, 83, 100, 103, 105, 110, 112, 115, 120, 124, 133, 140, 150.

RaiderOfLostVoodoo wrote on 2022-02-16, 16:30:

Question:
Are there any BX boards with UDMA66 and 3 ISA (or more) ISA slots?

I don't know for sure, but if I were to bet, I'd say that no. UDMA66 440BX boards usually have one (maybe two at most?) ISA slots. 3 ISA slots was common on the early 440BX boards.

1 x PLCC-68 / 2 x PGA132 / 5 x Skt 3 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 14 x Slot 1 / 5 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 5800X3D
Backup PC: Core i7 7700k

Reply 22 of 200, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

UDMA66 440BX boards usually have one (maybe two at most?) ISA slots.

ASUS CUBX-E had revision with 2 ISA slots.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 23 of 200, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
The Serpent Rider wrote on 2022-02-16, 17:37:

ASUS CUBX-E had revision with 2 ISA slots.

Nice! That's definitely a keeper. 😀

1 x PLCC-68 / 2 x PGA132 / 5 x Skt 3 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 14 x Slot 1 / 5 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 5800X3D
Backup PC: Core i7 7700k

Reply 24 of 200, by Warlord

User metadata
Rank l33t
Rank
l33t
RaiderOfLostVoodoo wrote on 2022-02-16, 16:30:

Question:
Are there any BX boards with UDMA66 and 3 ISA (or more) ISA slots?

not that I'm aware of. all the UDMA66 boards have i've seen have 1 ISA slot. Edit I quess cubx had 2.

One thing that isn't obviuos is the IRQ sharing sceme on BX boards. If it matters to you PCI slots share IRQs with other PCI slots and onboard devices. It is posible that on a board for example a UDMA 66 controller will share and IRQ with PCI slit 4 for example. But also PCI slot 4 would share a IRQ with PCI slot 2.

Which could mean if you use UDMA controller both PCI slot 4 and 2 are useless if you are trying to avoid IRQ sharing. So if you force UDMA controller to use like IRQ 10 for example. Any card inserted in slots 4 and 2 will also be 10.

So sounds goot till you start getting into the weeds. They end of giving you features in exchange for less control.

Reply 25 of 200, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Everything connected to 440BX south bridge is working directly through PCI or ISA bus. So yeah, "integrated" UDMA controllers can be a pain to deal with.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 26 of 200, by mothergoose729

User metadata
Rank Oldbie
Rank
Oldbie
Doornkaat wrote on 2022-02-16, 09:24:
janih wrote on 2022-02-16, 09:08:
bloodem wrote on 2022-02-16, 07:48:

Yeah. I mean, I do love SS7 for what it is (I love it BECAUSE of all its quirks and bugs). But I imagine that I'm in a minority - I think many people are purchasing SS7 builds because they overestimate their potential. It's enough to look at the prices of that platform to understand that there is a bit of a misunderstanding going on ($200 motherboards, $100+ K6-3+ CPUs). Meanwhile, an Ezra-T sells for $20 and Gigabyte 6BXC boards can still be bought for < $40 😁. And such a build absolutely DESTROYS any SS7 platform, not only in price, but also in its performance, stability and speed flexibility. The only advantage that a SS7 build brings to the table is the fun/entertainment of trying to solve all its problems, identifying the right combo of hardware/drivers/tweaks to achieve stability and decent speed (but, again, that's probably not what the average Joe is looking for).

One extra problem is the required slotket adapter and low(?) voltage for using an s370 VIA C3 on a slot1 motherboard. You've made an VRM mod for the 6BXC, is it needed for VIA C3 or do motherboards support the required voltages? I mean that is an extra hurdle compared to SS7 where you can just drop the cpu (=K6-3 for example) straight in 😀

There are plenty Slot 1 mobos that have a compatible VRM in their original/factory state. There are plenty native S370 boards that support VIA C3 out of the box as well. Probably lots more than there are Super7 boards even.😄 So no need for Slotkets or VRM mods just to run a VIA C3.

I think Bloodem chose a Slot 1 board simply so he could compare Katmai Pentium IIIs (that only come as Slot 1) to VIA C3s on the same board.

One nice think about the gigabyte 6xbc rev 2.0 board is that you can use third party tools to adjust the FSB in software. With my board I can hit any speed point I like without having to enter the bios.

Reply 27 of 200, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie

NEW VIDEO (VIA Nehemiah C3 Benchmark): https://www.youtube.com/watch?v=BDSClAu9boo

TEST SYSTEM REMINDER:
MB: Gigabyte GA-6BXC rev 1.9 (VRM mod)
CPU: VIA C3 Nehemiah 1.2 GHz (OC @ 1.46 GHz / FSB133 & 11.0 multi)
RAM: 3 x 128 MB SAMSUNG SDRAM PC133
VIDEO: Asus V7700Ti GeForce 2 Ti (OC @ GeForce 2 Ultra clocks)
SOUND: Creative Sound Blaster Live 5.1 SB0220
HDD: Seagate 40 GB IDE/PATA

This particular CPU, at its default voltage, handles a 1.53 GHz overclock just fine (which can be seen at the begining of the video), but I've decided to benchmark at the more common 1.46 GHz overclock, which should be easily achievable by all (or most) Nehemiah CPUs. All Nehemiah CPUs that I tried can do 1.46 GHz (some might require a very small voltage bump). The CPU that I'm using in the video can do 1.53 GHz at default voltage with a cheap slotket.
And, yes, this time I'm using a generic (FastFame?) slotket, which, as mentioned before, works extremely well (but has no manual voltage settings - not a problem in this case).

I'll let you decide how fast you think it is (it's pretty fast!) and how it compares to other CPUs, but more benchmarks will follow using the same system configuration (and other platforms like the SS7), so that we can (maybe) finally put to rest some of the constant debates that appear on this forum. 😁
I will point out that this CPU is not as flexible as the Ezra-T when it comes to obtaining 486DX-33/486DX2-66/486DX4-100 speeds. You can pretty much hit any speed point for 286 / 386 CPUs, you can also achieve the speed of a 486SX-25, but after that there is a huge gap until somewhere around a Pentium 100 - 133. It's debatable how important this is (in my particular case, it's really not) - a slow 386 / fast 386 / slow 486 and Pentium speed is all I need to play all of my favorite (speed sensitive) games. 😀

I'll edit the first post and point to each comment that contains a new video, to keep it simple for everyone.

And... another pic with the new CPU (well, the new cooler) and slotket (and the DIY "lock mechanism" 😅 )

Attachments

  • IMG_2968.jpeg
    Filename
    IMG_2968.jpeg
    File size
    653.42 KiB
    Views
    1567 views
    File license
    Fair use/fair dealing exception
Last edited by bloodem on 2022-02-17, 14:52. Edited 1 time in total.

1 x PLCC-68 / 2 x PGA132 / 5 x Skt 3 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 14 x Slot 1 / 5 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 5800X3D
Backup PC: Core i7 7700k

Reply 30 of 200, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie

Glad you guys like it. 😀

1 x PLCC-68 / 2 x PGA132 / 5 x Skt 3 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 14 x Slot 1 / 5 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 5800X3D
Backup PC: Core i7 7700k

Reply 31 of 200, by rasz_pl

User metadata
Rank l33t
Rank
l33t

2003, so C3 was released at the same time Intel sold Northwood based Celerons "2.60 GHz is priced at $103; the 2.50 GHz is priced at $89.", not to mention already on the market for a year tualatin Celeron 1400. From other tests on the web C3 was 2-3x slower than Coppermine Celerons at same MHz. Something tells me VIA wasnt selling them for $30.

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

Reply 33 of 200, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
mothergoose729 wrote on 2022-02-17, 18:27:

I am curious, what is the VRM mod for the gigabyte board? I think I can only hit 10x multi on my board and nehemiah.

The VRM mod is only needed for GA-6BXC boards older than rev 2.0 (the board I'm testing with is a rev 1.9 - which doesn't support Coppermines/Ezra-T/Nehemiah CPUs out of the box). It has nothing to do with multipliers.

As I explained in an earlier post, the mod consists in replacing the existing VRM chip (HIP6004ACB) on the rev 1.9 board, which only supports 1.8V or above, with a HIP6004BCB chip that supports voltages of 1.3V or above.

I also have 4 x rev 2.0 6BXC boards, but I specifically chose a rev 1.9 version to prove that it works without issues. 😀

1 x PLCC-68 / 2 x PGA132 / 5 x Skt 3 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 14 x Slot 1 / 5 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 5800X3D
Backup PC: Core i7 7700k

Reply 35 of 200, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
mothergoose729 wrote on 2022-02-17, 18:36:

ah ok, so just silicon lottery then. Thanks!

Oh, you mean it's not stable when you go past 10x?
That would be weird, I have 7 CPUs (which I bought from different sellers/on different occasions), and they all can do 1.46 GHz.

This is why I specifically tested at 1.46 GHz (even though this particular CPU works fine at 1.53 GHz), because it seemed to me like most Nehemiah chips should be able to easily reach this frequency (although some might need a small voltage increase).
So, either I was extremely lucky and managed to find 7 CPUs that are all excellent overclockers, or you are extremely unlucky 😁

Have you tested with different slotkets and maybe a different motherboard? Maybe there's more to it than just bad luck with the silicon lottery.

1 x PLCC-68 / 2 x PGA132 / 5 x Skt 3 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 14 x Slot 1 / 5 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 5800X3D
Backup PC: Core i7 7700k

Reply 36 of 200, by mothergoose729

User metadata
Rank Oldbie
Rank
Oldbie
bloodem wrote on 2022-02-17, 18:40:
Oh, you mean it's not stable when you go past 10x? That would be weird, I have 7 CPUs (which I bought from different sellers/on […]
Show full quote
mothergoose729 wrote on 2022-02-17, 18:36:

ah ok, so just silicon lottery then. Thanks!

Oh, you mean it's not stable when you go past 10x?
That would be weird, I have 7 CPUs (which I bought from different sellers/on different occasions), and they all can do 1.46 GHz.

This is why I specifically tested at 1.46 GHz (even though this particular CPU works fine at 1.53 GHz), because it seemed to me like most Nehemiah chips should be able to easily reach this frequency (although some might need a small voltage increase).
So, either I was extremely lucky and managed to find 7 CPUs that are all excellent overclockers, or you are extremely unlucky 😁

Have you tested with different slotkets and maybe a different motherboard? Maybe there's more to it than just bad luck with the silicon lottery.

You might be right on that, maybe the voltage on the slotket could be played with. For me, 10x multi is pretty good so I haven't felt that compelled to push it 😁.

Reply 37 of 200, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++
rasz_pl wrote:

2003, so C3 was released at the same time Intel sold Northwood based Celerons "2.60 GHz is priced at $103; the 2.50 GHz is priced at $89.", not to mention already on the market for a year tualatin Celeron 1400. From other tests on the web C3 was 2-3x slower than Coppermine Celerons at same MHz. Something tells me VIA wasnt selling them for $30.

They were primarily used for embedded systems with very low TDP, usually soldered on motherboard. CPGA models were more of an afterthought, probably as an upgrade path for old systems.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 38 of 200, by rasz_pl

User metadata
Rank l33t
Rank
l33t
The Serpent Rider wrote on 2022-02-17, 21:04:
rasz_pl wrote:

2003, so C3 was released at the same time Intel sold Northwood based Celerons "2.60 GHz is priced at $103; the 2.50 GHz is priced at $89.", not to mention already on the market for a year tualatin Celeron 1400. From other tests on the web C3 was 2-3x slower than Coppermine Celerons at same MHz. Something tells me VIA wasnt selling them for $30.

They were primarily used for embedded systems with very low TDP, usually soldered on motherboard. CPGA models were more of an afterthought, probably as an upgrade path for old systems.

Im super curious about blooden comparison testing. I have a suspicion Coppermine Celeron downclocked to 400-500MHz might be on par speed and TDP wise.

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

Reply 39 of 200, by bloodem

User metadata
Rank Oldbie
Rank
Oldbie
rasz_pl wrote on 2022-02-18, 03:10:

Im super curious about blooden comparison testing. I have a suspicion Coppermine Celeron downclocked to 400-500MHz might be on par speed and TDP wise.

I'm guessing you're referring to the Ezra-T, because the Nehemiah is obviously much faster than that. 😀

Well, it's definitely something worth exploring (and I do have a 600 MHz Celeron SL3W8, which I've never tested - I hope it works 😁 ). My early estimate is that it should be slower than the Ezra-T, because of the 66 MHz FSB (which becomes a severe bottleneck).
The Celeron 600 would definitely be slower than the Pentium 3 "Katmai" 600 (which I will try and test today, before I leave for a short one week vacation). Even if the Celeron has 128 KB of on-die L2 cache, its 66 MHz FSB is most likely enough to keep it well below the Katmai.

Now, I'm not sure what "downclocked" means to you. The Coppermine-128 Celeron has a locked multiplier, and the lower end parts run at 66 MHz FSB (not much wiggle room there).
So you are probably talking about downclocking a Celeron 800 (which runs at FSB 100) and running it at 66 MHz? That would basically give you a 533 MHz Celeron, which would be a bit slower than the 600 MHz Celeron I mentioned earlier, so might as well test that one instead.
Having said that, there's no way of getting a Coppermine-128 down to 400 - 500 MHz, other than lowering the FSB even further, down to 50 MHz. 😁 At that point it would get crushed by the Ezra-T.

The 600 MHz Duron, on the other hand, is most likely faster than the Ezra-T.

1 x PLCC-68 / 2 x PGA132 / 5 x Skt 3 / 9 x Skt 7 / 12 x SS7 / 1 x Skt 8 / 14 x Slot 1 / 5 x Slot A
5 x Skt 370 / 8 x Skt A / 2 x Skt 478 / 2 x Skt 754 / 3 x Skt 939 / 7 x LGA775 / 1 x LGA1155
Current PC: Ryzen 7 5800X3D
Backup PC: Core i7 7700k