VOGONS


VIA C3 Nehemiah vs. Coppermine(-128) and Tualatin

Topic actions

First post, by Kamerat

User metadata
Rank Oldbie
Rank
Oldbie

I wanted to know how the VIA C3 Nehemiah performed against a Pentium III so I did a benchmark run with a couple of CPUs. One reason to choose the Nehemiah are the number of options for slowdown (compared to Pentium II/III CPUs) if you also want to run speed sensetive DOS games. Note that none of the CPUs that I benchmarked on the Azza PT-6IBT are official supported, but on the positive side the board support voltages as low as 1,3V.

Test setup:
Azza PT-6IBT (i440BX, Slot 1, Baby AT)
VIA C3 "Nehemiah" 1.2A (on MS6905 MASTER)
Intel Pentium III-S "Tualatin" 1400 (modified, on MS6905 MASTER)
Intel Celeron "Coppermine-128" 1100 (on MS6905 MASTER)
Intel Pentium III "Coppermine" 866 (on MS6905 MASTER)
Intel Pentium III "Coppermine" 800EB (Slot 1)
Intel Pentium III "Coppermine" 800 (on Asus S370-DL)
256MB PC133 3-3-3-5
nVidia GeForce FX5200 128MB 64bit
Western Digital WD200 20GB

Windows 98SE
DirectX 8.1b
ForceWare 71.84
SetMul 1.2
SoftFSB 1.7-g1

Chart:

nehemiah-comparison.png
Filename
nehemiah-comparison.png
File size
73.64 KiB
Views
8510 views
File license
Fair use/fair dealing exception

Note the lack of the Nature test in 3DMark2001, I actually forgot that the FX5200 would run it just fine so I disabled it (I started to bench with a GeForce2 MX but it wouldn't cooperate). 😜

Conclusion:
For 100MHz bus a C3 at 1500MHz compares to a Pentium III at 700-800MHz in games.
For 133MHz bus a C3 at 1533MHz compares to a Pentium III at 733-800MHz in games.
A C3 at 1533MHz actually beats a Pentium III-S 1400MHz in Doom, the game seem to like high FSB very much.
Celeron "Coppermine-128" isn't very fast in Unreal Turnament.
I think the VIA C3 "Nehemiah" would do just fine with games up to 2000/2001.

Attachments

  • Filename
    nehemiah-comparison.7z
    File size
    27.77 KiB
    Downloads
    230 downloads
    File comment
    The chart in the original .ODS format
    File license
    Fair use/fair dealing exception

DOS Sound Blaster compatibility: PCI sound cards vs. PCI chipsets
YouTube channel

Reply 1 of 71, by j^aws

User metadata
Rank Oldbie
Rank
Oldbie

Pretty much my findings: Nehemiahs clock halved is roughly equivalent to the clock of its Pentium III counterpart. Especially for FP-heavy tasks. Not so much for integer-heavy tasks, though.

Reply 2 of 71, by clueless1

User metadata
Rank l33t
Rank
l33t

Nice! 😀 Thank you for sharing your results. I know that took some time to run all those benches. I agree with your assessment that the Nehemiah would make a good win98 PC for games up to 2001, with the added benefit of being more easily slowed down in controlled steps to run speed sensitive games.

The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks

Reply 3 of 71, by gerwin

User metadata
Rank l33t
Rank
l33t

Well done, I think it was about time to get some fair coverage for the VIA C3 Nehemiah. Past year there were a few reports here about it not working properly, even on VIA mainboards. Which is kinda strange since on most of the i440BX boards I have it runs like a charm. Also the Nehamiah gets a bad reputation from the poorly performing VIA C3 Samuel(2) and Ezra(-T). Yet despite, or thanks to, that poor performance the Ezra-T is still very interesting for retro gaming.

So I want to add some overdue compatibility tests I did today with this VIA C3 Nehemiah:

Working fine at 9.0x133=1200MHz, Quake 1 in 640x480 gives about 70 FPS on all these boards (write combining enabled), Speedsys gives +-800 points:
i440BX Slot 1 Gigabyte GA-6BXC Rev 2.0 - MS6905 Master v2.3 Slotket or Slot-T, Powerleap BIOS Award 6.00PG.
i440BX Slot 1 SOYO 6BA+III - MS6905 Master v2.3 Slotket, BIOS BA+III_2BA6.
i440BX Slot 1 Asus P2B v1.04 - (With VRM upgrade) MS6905 Master v2.3 Slotket, BIOS 1014 Beta 003 with ROM.by Tualatin patch.
i440BX Slot 1 Aopen AX6BC v1.4 - MS6905 Master v2.3 Slotket, BIOS R2.59.
Apollo133A S370 Chaintech CT6AJA4 - IIRC with Lin-Lin adapter.

Did not work:
i440BX Slot 1 Shuttle HOT-661 v1.1 - (With VRM upgrade) MS6905 Master v2.3 Slotket, Got only up to VIA C3 Samuel 2 running, this is not a BIOS limitation.
i440BX Slot 1 Intel SR440BX - MS6905 Master v2.3 Slotket, None of the C3 CPUs can boot, probably a BIOS restriction.

--> ISA Soundcard Overview // Doom MBF 2.04 // SetMul

Reply 4 of 71, by j^aws

User metadata
Rank Oldbie
Rank
Oldbie

^^ Have you tried different Slotkets/ Slockets to get those boards to work? Or messed with their jumper settings? I've gotten various Slotkets/ Slockets from different manufacturers to work, but some don't like certain boards.

Regarding VIA C3 Ezra-T, it's like the S3 ViRGE of retro computing: Ridiculed for its deceleration capabilities at launch, but like a fine wine, appreciated with age.

Reply 5 of 71, by gerwin

User metadata
Rank l33t
Rank
l33t
j^aws wrote:

^^ Have you tried different Slotkets/ Slockets to get those boards to work? Or messed with their jumper settings? I've gotten various Slotkets/ Slockets from different manufacturers to work, but some don't like certain boards.

When a board boots with the Nehemiah, it kinda implies that there are different combinations of Slotkets that should work. I kinda stuck to the MS6905 Master Slotket because it is still plentiful on ebay.de, and because it has a voltage clamp chip. UpgradeWare Slot-T works just as well. A cheap Super Slocket III with lin-lin adapter on top works too. I don't have time now to record all the results of different jumper settings.

I am not finished with the HOT-661. I am intrigued why it does not work, even though it is so similar to the GA-6BXC that it works when I use the GA-6BXC BIOS.

j^aws wrote:

Regarding VIA C3 Ezra-T, it's like the S3 ViRGE of retro computing: Ridiculed for its deceleration capabilities at launch, but like a fine wine, appreciated with age.

Well said.

--> ISA Soundcard Overview // Doom MBF 2.04 // SetMul

Reply 7 of 71, by gerwin

User metadata
Rank l33t
Rank
l33t

Intel SR440BX so far does not cooperate. Maybe the BIOS does only allows intel vendor strings, just like it does not allow 133MHz FSB intel processors. I expect SE440BX-2 to behave the same.

--> ISA Soundcard Overview // Doom MBF 2.04 // SetMul

Reply 8 of 71, by matze79

User metadata
Rank l33t
Rank
l33t

C3's FPU works only with half CPU Clock.

https://www.retrokits.de - blog, retro projects, hdd clicker, diy soundcards etc
https://www.retroianer.de - german retro computer board

Reply 9 of 71, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++
gerwin wrote:

Intel SR440BX so far does not cooperate. Maybe the BIOS does only allows intel vendor strings, just like it does not allow 133MHz FSB intel processors. I expect SE440BX-2 to behave the same.

Hmm... anyone familiar with modding Phoenix bioses?

All hail the Great Capacitor Brand Finder

Reply 10 of 71, by gerwin

User metadata
Rank l33t
Rank
l33t
matze79 wrote:

C3's FPU works only with half CPU Clock.

I have read that Nehemiah core introduced a full speed FPU. Normally one can do a benchmark comparison to verify such, but the preceding cores like Ezra are so peculiar...

--> ISA Soundcard Overview // Doom MBF 2.04 // SetMul

Reply 11 of 71, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

Do any of your working motherboards use a Phoenix BIOS, or are they all Award or AMI? With luck, we can port across a CPU support module for the Intel board.

All hail the Great Capacitor Brand Finder

Reply 12 of 71, by gerwin

User metadata
Rank l33t
Rank
l33t

The Soyo 6BA+III has a 'Phoenix Technologies AwardBIOS'. It identifies the Nehemiah as a 'VIA CyrixIII 1000MHz'.

Though can you risk flashing a modded BIOS to the intel SE440BX/SR440BX? These boards have their EEPROM soldered to the board. So in case a new BIOS fails to boot: No way to revert back to a working BIOS. You would be left with a bricked motherboard, unless there is some recovery method I don't know about.

For tinkering these intel designed Motherboards are the worst starting point. ( A few simple hardware mods are still possible )

--> ISA Soundcard Overview // Doom MBF 2.04 // SetMul

Reply 13 of 71, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++

Unusual form factor for the flash chip, so you can't even socket it easily.

Poop.

All hail the Great Capacitor Brand Finder

Reply 14 of 71, by j^aws

User metadata
Rank Oldbie
Rank
Oldbie
matze79 wrote:

C3's FPU works only with half CPU Clock.

DSC_0325.JPG
Filename
DSC_0325.JPG
File size
3.2 MiB
Views
7951 views
File license
Fair use/fair dealing exception
DSC_0326.JPG
Filename
DSC_0326.JPG
File size
3.2 MiB
Views
7951 views
File license
Fair use/fair dealing exception

You can see above that NSSI reports for an Ezra-T that both processor and co-processor (FPU) are running at the same speed of 1072 MHz (this is slightly off anyway, but the point being the same speed). The later Nehemiah should also be running the same. Even then, these FPUs are still fairly weak. Nehemiah being stronger than Ezra-T.

Reply 15 of 71, by Kamerat

User metadata
Rank Oldbie
Rank
Oldbie

Thanks for the inputs people of Vogons. 😊

j^aws wrote:

Pretty much my findings: Nehemiahs clock halved is roughly equivalent to the clock of its Pentium III counterpart. Especially for FP-heavy tasks. Not so much for integer-heavy tasks, though.

I was afraid that it would perform even worse after I saw the bad results for SuperPI.

clueless1 wrote:

Nice! 😀 Thank you for sharing your results. I know that took some time to run all those benches. I agree with your assessment that the Nehemiah would make a good win98 PC for games up to 2001, with the added benefit of being more easily slowed down in controlled steps to run speed sensitive games.

One problem with the slowdown is the large gap in performance between running 5,5x66MHz with level 2 cache and branch prediction disabled and running 11,5x133MHz with L1 I-cache disabled, at least Decent ran fine with the first setting. Trying multipliers lower than 5,5 when running 66MHz FSB makes the system crash on the Azza board.

gerwin wrote:

Well done, I think it was about time to get some fair coverage for the VIA C3 Nehemiah. Past year there were a few reports here about it not working properly, even on VIA mainboards. Which is kinda strange since on most of the i440BX boards I have it runs like a charm. Also the Nehamiah gets a bad reputation from the poorly performing VIA C3 Samuel(2) and Ezra(-T). Yet despite, or thanks to, that poor performance the Ezra-T is still very interesting for retro gaming.

So I want to add some overdue compatibility tests I did today with this VIA C3 Nehemiah:

Thanks for the list and the great SetMul tool. I also had the Nehemiah working on the Asus TUSC (SiS 630ET, socket 370, latest beta BIOS), but it's not official supported on this board either. On the TUSC I were able to run the lowest multiplier (4x) even with 66MHz FSB. I might get an Ezra-T to play with just for the fun of it. 😀

DOS Sound Blaster compatibility: PCI sound cards vs. PCI chipsets
YouTube channel

Reply 16 of 71, by Kamerat

User metadata
Rank Oldbie
Rank
Oldbie
gdjacobs wrote:

Does the SE440BX-2 work with a C3?

gerwin wrote:

Intel SR440BX so far does not cooperate. Maybe the BIOS does only allows intel vendor strings, just like it does not allow 133MHz FSB intel processors. I expect SE440BX-2 to behave the same.

Just had to try it, running the C3 "Nehemiah" on my SE440BX-2 right now. I'm using the P13 BIOS.

DOS Sound Blaster compatibility: PCI sound cards vs. PCI chipsets
YouTube channel

Reply 17 of 71, by Tetrium

User metadata
Rank l33t++
Rank
l33t++
Kamerat wrote:
gdjacobs wrote:

Does the SE440BX-2 work with a C3?

gerwin wrote:

Intel SR440BX so far does not cooperate. Maybe the BIOS does only allows intel vendor strings, just like it does not allow 133MHz FSB intel processors. I expect SE440BX-2 to behave the same.

Just had to try it, running the C3 "Nehemiah" on my SE440BX-2 right now. I'm using the P13 BIOS.

Perhaps because its BIOS only blocks (mostly Intel) CPUs that already existed or were planned at the time your BIOS got released?
If a newer BIOS blocks more CPUs, then one 'could' assume that these CPUs are specifically blocked purposefully (but one cannot specifically block something purposefully that didn't exist yet). This is only an assumption though.

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

Reply 18 of 71, by Kamerat

User metadata
Rank Oldbie
Rank
Oldbie
Tetrium wrote:

Perhaps because its BIOS only blocks (mostly Intel) CPUs that already existed or were planned at the time your BIOS got released?
If a newer BIOS blocks more CPUs, then one 'could' assume that these CPUs are specifically blocked purposefully (but one cannot specifically block something purposefully that didn't exist yet). This is only an assumption though.

For this board it might be the case. A while ago I downgraded the BIOS because it wouldn't run Coppermines, I got picture but it displayed an error message and wouldn't let me continue. I can also boot my modified Pentium III-S on that board. Both the Nehemiah and Tualatin got detected as Pentium Pros by the BIOS, but the Pentium III 800 "Coppermine" still got detected as a Pentium III IIRC.

DOS Sound Blaster compatibility: PCI sound cards vs. PCI chipsets
YouTube channel

Reply 19 of 71, by gdjacobs

User metadata
Rank l33t++
Rank
l33t++
Kamerat wrote:

Just had to try it, running the C3 "Nehemiah" on my SE440BX-2 right now. I'm using the P13 BIOS.

Sweet!

All hail the Great Capacitor Brand Finder