shevalier wrote on 2026-07-06, 04:35:

RAM chips from Hynix? In my opinion, their PC133 chips aren't good for overclocking. To be successful, you need a motherboard th […]
Show full quote

Repo Man11 wrote on 2026-07-05, 17:26:

Same as before, the highest 3d 2000 score is with the Ti 4600.

Worth mentioning is the memory, Kingston KVR133X64C3 - the 512 meg stick I have handles 144 MHz with tight timings (though CAS 3) at default voltage. Most of the generic PC133 I have falls short of this.

RAM chips from Hynix?
In my opinion, their PC133 chips aren't good for overclocking.
To be successful, you need a motherboard that supports 166/166 (FSB/RAM) and memory that can run 5-2-2-2 at that frequency.
With high-end Pentium 3 processors, increasing the core frequency no longer yields significant gains; the FSB becomes the limiting factor.

This is likely why 1.4–1.5 GHz is the official limit for these cores.

The chips are labeled Nanya. I thought I'd mention it because this stick is the only one I have that will do 144 with tight timings on a board that offers no ability to increase the RAM voltage; looking at Ebay, this Kingston brand RAM is still available at reasonable prices. I've no doubt that better can be had.

Repo Man11 wrote on 2026-07-06, 13:15:

I've no doubt that better can be had.

There was a thread on the forum about selecting and overclocking PC133 memory.
The author managed to reach 200 MHz, but the number of memory modules he had purchased and tested was impressive.

Overclocking PC133 SDR RAM: What is the best SDR memory?

supercordo wrote on 2026-06-20, 19:09:

Heres a good post on what ram to look for, for a higher FSB.

Overclocking PC133 SDR RAM: What is the best SDR memory?

This is the 7th post here.

Ooh, it's a Pentium III thread! I know this!

Pentium III-S 1400 @ 1628 MHz, FSB-155
QDI Advance 12T - Via Apollo Pro 266T chipset
2GB DDR-310, 2-2-2-5
EVGA GeForce 6800 GT @ 385/540
X-Fi Platinum
XP SP3

3DMark 2001SE

3DMark 2000 - Interestingly my R9800 Pro outperforms the 6800 GT in this benchmark; IIRC it gets around 12700. Likely because the drivers are leaner, and 3DMark 2000 doesn't have a GPU-heavy Nature equivalent to inflate the score.

Anyone want me to hold their beer?

It was only a matter of time before someone with the right combo showed up.

I'll get a 3D Mark 99 score on the board.

supercordo wrote on 2026-07-07, 01:40:

supercordo wrote on 2026-06-20, 19:09:

Heres a good post on what ram to look for, for a higher FSB.

Overclocking PC133 SDR RAM: What is the best SDR memory?

This is the 7th post here.

Sorry, I didn't see it.

Mendocino needs some more love! - so I've bashed together something. Haven't been able to get 3DMark 2001SE to install and 3DMark 99MAX isn't running stable, but have some solid reproducible results for 3DMark 2000 and CPUMark 99.

Abit BP6, 2x Celeron 333@500, 256MB PC133 (@100) CL2, Gf4Ti4200; Win98SE, Detonator 45.23

3DMark 2000 score 2823

CPUmark 99 score 40.8

Of course anyone with a good single-CPU board should be able to blow this away, that second CPU isn't doing anything sensible here, just reducing the max attainable speeds. I tried 104MHz FSB, but that hardly managed to get to desktop, even after raising VCore to 2.2V.

I've just repaired another MSI MS-6163 Ver 2 (Pro) board and, as I was testing it, I remembered this thread.

Benchmarked it with a modded Tualatin 1.4 GHz CPU + MSI MS-6905 "Master" ver 2.3 slotket, even if all capacitors are toast (it will receive a full recap over the next few days, using Panasonic low ESR capacitors).

But, even so, the board is still 100% stable @ FSB 142. Anything beyond that results in a freeze in 3DMark2000 (it will be interesting to see if this changes after the recap - will report back).
The GPU I used for these tests is a Medion GeForce 3 Ti 200, which is permanently overclocked to somewhere between a pure GeForce 3 & GeForce 3 Ti 500 clocks (230 MHz core / 460 MHz memory).

I hope it's OK that I used CrystalCPUID and Aida64 for the screenshots (I don't have CPU-Z or Everest on my Win98 machines).

bloodem wrote on 2026-07-10, 17:41:

Benchmarked it with a modded Tualatin 1.4 GHz CPU + MSI MS-6905 "Master" ver 2.3 slotket, even if all capacitors are toast (it will receive a full recap over the next few days, using Panasonic low ESR capacitors).

But, even so, the board is still 100% stable @ FSB 142. Anything beyond that results in a freeze in 3DMark2000 (it will be interesting to see if this changes after the recap - will report back).

Well, I'm back after fully recapping the board with Panasonic FR caps... Unfortunately, the overclock headroom on this particular board has not improved at all, so the max achievable (stable) FSB speed remains 142MHz.
However, I now tested it with a different video card, a GeForce 4 Ti 4200 (overclocked somewhere between a Ti 4400 and a Ti 4600), and this gave me an additional score boost (which I didn't really expect, since, past a certain point, these tests become mostly CPU-bound). Once again, the 440BX chipset stands tall. 😁

Bazinga!!!

I tried again with a Tualatin in my 7VMA-B. I used a Lin Lin adapter this time, and it works great (the Tualatin that I purchased with a preinstalled interposer wouldn't work in the board because it didn't play nice with this board's socket but it works well with my Tyan board). The CMOS settings jump from 144 FSB to 155 (!), so after having such good luck using CPU FSB with the Tyan board, I tried it with this one. It works well and there were many steps in between 144 and 155. I managed to hit 149 with a slight CPU core voltage increase. Though I'm happy with the result, it also serves to show the insurmountable gap between the Via chipset and the BX. With a slightly higher FSB, a slightly higher overall clock speed, and a better video card, Bloodem's rig still bests it by almost 1,400 points! Still it's pretty good, and for a board that likely spent its life with something like a 733 MHz Coppermine, the performance is fairly impressive

I'd love to try a 1.4 and see if I could run it at 1,565 MHz but they are silly expensive; I paid just over $20.00 for this 1.26 GHz chip but all of the 1.4s I've seen listed are over $100.00. So I'll sit tight on this.

bloodem wrote on 2026-07-16, 13:08:

... after fully recapping the MSI MS-6163 Ver 2 (Pro) with Panasonic FR caps ...

If you don't mind me asking, which FRs did you select for CPU VRM output?

TELVM wrote on 2026-08-05, 05:25:

If you don't mind me asking, which FRs did you select for CPU VRM output?

Don't mind it at all. 😀
If I remember correctly, those were 1000uF / 6.3V caps originally, and I replaced them with 1000uF / 10V Panasonic FR caps: https://www.tme.eu/en/details/eeufr1a102l/tht … tors/panasonic/

bloodem wrote on 2026-08-05, 09:01:

... I replaced them with 1000uF / 10V Panasonic FR caps ...

Thanks.

Again If you don't mind a constructive comment, those CPU VRM output caps will never have to handle a voltage higher than 2V, so 6.3V rated ones offer ample of headroom (even 4V rated are perfectly fine here).

The advantage of a lower (but still way safe) voltage rating is a lower ESR, which is what CPU VRM output caps love the most 😀 .

6x Panny FR 10V 1000uF 8x15mm (ESR= 41) => Combined: 41/6 = 6.83

6x Panny FR 6.3V 1500uF 8x20mm (ESR= 30) => Combined 30/6 = 5

(The increase in capacitance isn't really needed, but wouldn't harm).

They indeed can be polymodded (btw sorry y'all for being MIA, life got in the way) as I owned TWO of those (v1 and v2) both polymodded.

One has used 1200uF 4v polys to acommodate any Celeron overclock down the road (as in, any voltage bump needed to have at least a 2v headroom).
The other ran 820uF 2.5v polys.

1500uF is already too overkill factory - these will run on 820uF easily.

As for 133FSB issues I've had across all of them - apparently clockgen issue. Have been able to replicate the issue on Soyo 6BA series boards as well.

TELVM wrote on 2026-08-05, 11:03:

Again If you don't mind a constructive comment, those CPU VRM output caps will never have to handle a voltage higher than 2V, so […]
Show full quote

bloodem wrote on 2026-08-05, 09:01:

... I replaced them with 1000uF / 10V Panasonic FR caps ...

Again If you don't mind a constructive comment, those CPU VRM output caps will never have to handle a voltage higher than 2V, so 6.3V rated ones offer ample of headroom (even 4V rated are perfectly fine here).

The advantage of a lower (but still way safe) voltage rating is a lower ESR, which is what CPU VRM output caps love the most 😀 .

6x Panny FR 10V 1000uF 8x15mm (ESR= 41) => Combined: 41/6 = 6.83

6x Panny FR 6.3V 1500uF 8x20mm (ESR= 30) => Combined 30/6 = 5

(The increase in capacitance isn't really needed, but wouldn't harm).

Well, the lower ESR is not related to the lower voltage rating, it’s purely correlated with the physical case size, as mentioned in the Panasonic documentation.
In your example, the 1500uF / 6.3V cap is taller (20 mm), which is why it has lower ESR.

I usually prefer to use shorter caps, and I always try to match the original size as much as possible, just for aesthetic purposes. Plus, I didn’t have any 6.3V caps left. 😀
Also, these old boards don’t really need very low ESR, I’m sure they’d work perfectly with general purpose caps as well. For example, this board that I just fixed had faulty caps with out-of-spec ESR ratings of 5 or even 10+ ohms, and it was still working perfectly, even in this state.

PcBytes wrote on 2026-08-05, 14:54:

They indeed can be polymodded (btw sorry y'all for being MIA, life got in the way) as I owned TWO of those (v1 and v2) both poly […]
Show full quote

They indeed can be polymodded (btw sorry y'all for being MIA, life got in the way) as I owned TWO of those (v1 and v2) both polymodded.

One has used 1200uF 4v polys to acommodate any Celeron overclock down the road (as in, any voltage bump needed to have at least a 2v headroom).
The other ran 820uF 2.5v polys.

1500uF is already too overkill factory - these will run on 820uF easily.

As for 133FSB issues I've had across all of them - apparently clockgen issue. Have been able to replicate the issue on Soyo 6BA series boards as well.

I think I mentioned this before in a different discussion, but I did not encounter the instability you mention on any of my MS-6163 boards, and I’ve had them and tested them for many years.
Yes, in my experience, ver. 2 does tend to be a little more fussy with certain older video cards @ FSB133 (potentially resulting in some artifacts), but it works flawlessly with any GeForce 3 & GeForce 4 card I ever tried.

Hell, even now I tested for a few weeks this latest one that I repaired (using both a Tualatin P3 1.4 GHz and a Coppermine P3 1 GHz) and it was rock solid, even at 142 MHz FSB (at one point I even forgot I left it in a 3DMark2001 benchmark loop, and it ran like this for 6 hours or so).
So, I’m very curious to understand what combination of hardware (maybe software too?) causes the behavior you’re seeing.

bloodem wrote on 2026-08-06, 02:25:

... Well, the lower ESR is not related to the lower voltage rating, it’s purely correlated with the physical case size ...

Yep, the larger the lytic can size, the lower the ESR and the higher the ripple current.

But for the same can size, the higher the voltage rating, the lower the capacitance.

Panny FR 8x15mm 6.3V: 1200uF
Panny FR 8x15mm 10V: 1000uF
Panny FR 8x15mm 16V: 680uF
Panny FR 8x15mm 25V: 470uF

^ All same ESR & ripple.

bloodem wrote on 2026-08-06, 02:25:

... these old boards don’t really need very low ESR, I’m sure they’d work perfectly with general purpose caps as well ...

I'm afraid Intel disagrees.

Intel Slot 1 Processor Power Distribution Guidelines

This wonderful graph explains at a glance the relationship between capacitance and ESR for the CPU VRM output bank of caps:

^ General purpose lytics with high ESR fall beyond the right edge of the graph , where you need a humongous amount of combined capacitance to stay above the grey insufficiency area.

Low ESR lytics fall around the middle-right in the graph.

Very low ESR polys fall to the left in the graph, where you can still stay above the gray area even with a much reduced combined capacitance (5000~6000uF).

Notice how even with hypothetical magical caps of ESR= zero, we'd still need ~4750uF of combined capacitance (under the particular parameters Intel chose for this graph).

.

IMHO it's not a good idea to place high ESR general purpose caps at the CPU VRM output capacitor bank.

I have 2+1 (fried from testing) CUBX-L motherboards. They were pulled from the exact same IT department scrap box and are of identical revisions.
The dead one back in the day could run 1 GB of RAM and even a Radeon 9800 XT at a 133 MHz bus, while the other two top out at 124 MHz and won't run the 9800 at all. On top of that, one of them has been fully recapped with polymer capacitors, while the second one is completely untouched.

It all comes down to the individual motherboard when you try pushing it beyond official specs.

TELVM wrote on 2026-08-06, 06:55:

I'm afraid Intel disagrees. […]
Show full quote

I'm afraid Intel disagrees.

Intel Slot 1 Processor Power Distribution Guidelines

This wonderful graph explains at a glance the relationship between capacitance and ESR for the CPU VRM output bank of caps:

^ General purpose lytics with high ESR fall beyond the right edge of the graph , where you need a humongous amount of combined capacitance to stay above the grey insufficiency area.

Low ESR lytics fall around the middle-right in the graph.

Very low ESR polys fall to the left in the graph, where you can still stay above the gray area even with a much reduced combined capacitance (5000~6000uF).

Notice how even with hypothetical magical caps of ESR= zero, we'd still need ~4750uF of combined capacitance (under the particular parameters Intel chose for this graph).

.

IMHO it's not a good idea to place high ESR general purpose caps at the CPU VRM output capacitor bank.

I think you've misunderstood what I was saying. I know very well what Intel & AMD say, and of course low ESR caps will always be recommended for buck converters, which is why I myself use them. 😀
I'm just saying that, based on my experience, even if you go the bad route and use general purpose caps, the truth of the matter is that... nothing bad will happen.

As I said (and I'm sure many have had a similar experience), boards can sometimes work fine even with leaky caps that are WAY out of spec. Yes, the VRM will be under a lot of stress, and prolonged operation will certaninly increase the risk of permanent damage, but many times the board will not show any signs of instability (and, again, we're talking 200 to 300 times higher ESR than normal!). The only plausible explanation is that these circuits probably have far more tolerance than the design guidelines appear to suggest.

22 years ago, I used to work for a computer repair company which would only source cheap general purpose caps for all repairs. I can't even tell you how many recaps we did for all kinds of boards ranging from socket 7 all the way up to socket 939 & 478. We never once received any complaint... ever (and most of those motherboards came from corporate clients, some of whom were running very demanding software, such as ship design specific apps).
But, again, I'm not advocating for the use of general purpose caps, just stating my experience. 😀

shevalier wrote on 2026-08-06, 06:58:

I have 2+1 (fried from testing) CUBX-L motherboards. They were pulled from the exact same IT department scrap box and are of identical revisions.
The dead one back in the day could run 1 GB of RAM and even a Radeon 9800 XT at a 133 MHz bus, while the other two top out at 124 MHz and won't run the 9800 at all. On top of that, one of them has been fully recapped with polymer capacitors, while the second one is completely untouched.

It all comes down to the individual motherboard when you try pushing it beyond official specs.

Agreed. It's just that @PcBytes has (or had?) two such boards and I have... at least 5 that I remember right now, so it is kind of strange how different our experience is: all of my boards work (almost) fine @ 133 MHz, while all of his boards appear to have major issues.

I mean using 4 RAM sticks totaling 1 GB. With any 3 (of the ones I have), I can get 133 MHz with no issues at all. 🙁

Regarding polymer caps:
5 x 820 µF general-purpose ones will give you an ESR of right around 2 mΩ.
Works like a charm.
Ripple + noise is around 20–30 mV peak-to-peak.