I suspect that the Celeron would win most benchmarks at the higher clock speed (with the same FSB). The Celeron will lose out in certain tasks that saturate it's 256K of cache. Rare, though.

I had a 1300MHz Celeron (100MHz FSB) back in the day, and it was a beast!

Definitely interesting. You can make up for a lot of cache misses with 333mhz of additional clock speed.

See this Re: Tualatin Celeron vs Williamette Celeron

EDIT: That extra cache really makes a difference .

darry wrote on 2021-01-13, 21:08:

See this Re: Tualatin Celeron vs Williamette Celeron

EDIT: That extra cache really makes a difference .

cel _9.5 x 133 = 1266 MHz: 7506 3DMarks
p3 _9.5 x 133 = 1266 MHz: 7845 3DMarks

~4,5%

1733/1400 = ~24% frequency difference

My 1300 is B1 stepping and I'm sure it will work

You don't. Test it thoroughly for stability, then proceed. If you can't do that, then your best bet would be PIII-S. Also keep in mind that stable overclock may degrade over time, especially it it's borderline stable.

1733/1400 = ~24% frequency difference

CPU frequency won't make much dent, because high-end PIII's are really starved on slow FSB.

overdrive333 wrote on 2021-01-13, 21:45:

cel _9.5 x 133 = 1266 MHz: 7506 3DMarks p3 _9.5 x 133 = 1266 MHz: 7845 3DMarks […]
Show full quote

darry wrote on 2021-01-13, 21:08:

See this Re: Tualatin Celeron vs Williamette Celeron

EDIT: That extra cache really makes a difference .

cel _9.5 x 133 = 1266 MHz: 7506 3DMarks
p3 _9.5 x 133 = 1266 MHz: 7845 3DMarks

~4,5%

1733/1400 = ~24% frequency difference

What video card is that with ?

The Serpent Rider wrote on 2021-01-13, 21:55:

You don't. Test it thoroughly for stability, then proceed. If you can't do that, then your best bet would be PIII-S. Also keep i […]
Show full quote

My 1300 is B1 stepping and I'm sure it will work

You don't. Test it thoroughly for stability, then proceed. If you can't do that, then your best bet would be PIII-S. Also keep in mind that stable overclock may degrade over time, especially it it's borderline stable.

1733/1400 = ~24% frequency difference

CPU frequency won't make much dent, because high-end PIII's are really starved on slow FSB.

Agreed.

Would you be willing to benchmark both? I would be curious to see how they fair. I am not sure the question is totally settled just yet.

EDIT: Both CPUs benched at 100mhz FSB in addition to 133mhz FSB would also be interesting to see.

I actually tried this with a Celeron-1400 back in the day. Had it running at 1866/133, and at that speed it could match the stock-clocked PIII-S 1400 in some benchmarks. In others, the PIII was ahead. The Celeron also required some enormous voltage to get to 1866, and actually quit working after a few hours of benchmarking.

So yeah, definitely go with PIII-S 1400. It'll outperform the Celeron-1733 and run much cooler. Plus, most of them will happily do 1575/150 at stock voltage, should your motherboard accommodate.

Worth noting: I ran those tests on an old Asus TUV4X (Apollo 133) motherboard, which has lower memory performance than 440BX at equivalent bus clocks. That may have allowed the PIII-S and its ginormous cache to shine more brightly than it otherwise would have. 😀

Standard Def Steve wrote on 2021-01-13, 23:06:

I actually tried this with a Celeron-1400 back in the day. Had it running at 1866/133, and at that speed it could match the stock-clocked PIII-S 1400 in some benchmarks. In others, the PIII was ahead. The Celeron also required some enormous voltage to get to 1866, and actually quit working after a few hours of benchmarking.

So yeah, definitely go with PIII-S 1400. It'll outperform the Celeron-1733 and run much cooler. Plus, most of them will happily do 1575/150 at stock voltage, should your motherboard accommodate.

Worth noting: I ran those tests on an old Asus TUV4X (Apollo 133) motherboard, which has lower memory performance than 440BX at equivalent bus clocks. That may have allowed the PIII-S and its ginormous cache to shine more brightly than it otherwise would have. 😀

Oh, well there you have it 😁

I can confirm what Standard Def Steve wrote above.

Standard Def Steve wrote on 2021-01-13, 23:06:

I actually tried this with a Celeron-1400 back in the day. Had it running at 1866/133, and at that speed it could match the stock-clocked PIII-S 1400 in some benchmarks. In others, the PIII was ahead. The Celeron also required some enormous voltage to get to 1866, and actually quit working after a few hours of benchmarking.

But it sounds like you had a lot of fun doing it 🤣

My results from 3dmark 99:
Celeron 1400 - 17213 cpu marks
Celeron 1666 - 20585 cpu marks (c1200 @133 MHz fsb)
Tualatin 1400 - 21760 cpu marks

H3nrik V! wrote on 2021-01-15, 07:46:

Standard Def Steve wrote on 2021-01-13, 23:06:

I actually tried this with a Celeron-1400 back in the day. Had it running at 1866/133, and at that speed it could match the stock-clocked PIII-S 1400 in some benchmarks. In others, the PIII was ahead. The Celeron also required some enormous voltage to get to 1866, and actually quit working after a few hours of benchmarking.

But it sounds like you had a lot of fun doing it 🤣

Well, I was definitely excited when the thing posted and booted into Windows! I mean, a near 1.9GHz Tualatin! I was hoping for some Athlon XP, possibly even ULV Pentium M levels of performance! But then I ran the benchmarks and I was like: 🥱. And then it burned out and I was like: 😠