VOGONS


First post, by karakarga

User metadata
Rank Newbie
Rank
Newbie

Hello everyone,

As far as I know, there was a Pentium bug on first group of manufactured processors between 60 ~ 100 MHz. 120 MHz and up, does not have the FDIV bug! I have a complete list.
B5 stepping based Pentium 90 test without the bug, at this link https://www.youtube.com/watch?v=2OgP2WH5l_0 and there says https://forum.vcfed.org/index.php?threads/nee … fdiv-bug.18501/ all 63 & 83 MHz 486 overdrive processors are not affected!

Wikipedia also says: "The FDIV bug affects the 60 and 66 MHz Pentium P5 800 in stepping levels prior to D1, and the 75, 90, and 100 MHz Pentium P54C 600 in steppings prior to B5. Starting from 120 MHz (P54C and P54CQS) CPU's and up are, unaffected." https://en.wikipedia.org/wiki/Pentium_FDIV_bu … Affected_models

Pentium Overdrive for 486 systems --> Socket 2/3
Product Code -- Type -- Family -- Model -- Stepping -- Core stepping -- Clock rate -- FSB speed -- sSpec -- Version
PODP5V63 -- 1 -- 5 -- 3 -- 1 -- B1 -- 63 MHz -- 25 MHz -- SZ953 -- 1.0 (*) Not affected!
PODP5V63 -- 1 -- 5 -- 3 -- 1 -- B2 -- 63 MHz -- 25 MHz -- SZ990 -- 1.1 (*) Not affected!
PODP5V63 -- 1 -- 5 -- 3 -- 1 -- C0 -- 63 MHz -- 25 MHz -- SU013 -- 1.0 (*) Not affected!
PODP5V83 -- 1 -- 5 -- 3 -- 2 -- C0 -- 83 MHz -- 33 MHz -- SU014 -- 2.1 (*) Not affected!

Pentium P5 800 nm 5V --> Socket 4
Family -- Model -- Stepping -- Core stepping -- Clock rate -- FSB speed -- sSpec
5 -- 1 -- 1 -- B1 -- 60 MHz -- 60 MHz -- Q0352, Q0412, SX753
5 -- 1 -- 3 -- B1 -- 66 MHz -- 66 MHz -- Q0353, Q0413, SX754
5 -- 1 -- 5 -- C1 -- 60 MHz -- 60 MHz -- Q0466, SX835, SZ949
5 -- 1 -- 5 -- C1 -- 66 MHz -- 66 MHz -- Q0467, SX837, SZ950
5 -- 1 -- 5 -- D1 -- 60 MHz -- 60 MHz -- Q0625, SX948, SX974 (*) Not affected!
5 -- 1 -- 5 -- D1 -- 66 MHz -- 66 MHz -- 00626, 00627, SX949, SX950 (*) Not affected!

Pentium Overdrive for 586 systems --> Socket 4
Product Code -- Type -- Family -- Model -- Stepping -- Core stepping -- Clock rate -- FSB speed -- sSpec -- Version
PODP5V133 -- 0 -- 5 -- 1 -- aA -- tA0 -- 120/133 -- 60/66 -- SU082 -- 1.0 (*) Not affected!

Pentium P54C 600 nm 3.3V --> Socket 5/7
Family -- Model -- Stepping -- Core stepping -- Clock rate -- FSB speed -- sSpec
5 -- 2 -- 1 -- B1 -- 75 MHz -- 50 MHz -- Q0601
5 -- 2 -- 1 -- B1 -- 90 MHz -- 60 MHz -- Q0542, Q0613, Q0543, SX879, SX885, SX909, SX874
5 -- 2 -- 1 -- B1 -- 100 MHz -- 66 MHz -- Q0563, Q0587, Q0614, SX886, SX910
5 -- 2 -- 2 -- B3 -- 75 MHz -- 50 MHz -- Q0606, SX951
5 -- 2 -- 2 -- B3 -- 90 MHz -- 60 MHz -- Q0628, Q0611, Q0612, SX923, SX922, SX921, SX942, SX943, SX944, SZ951
5 -- 2 -- 2 -- B3 -- 100 MHz -- 66 MHz -- Q0677, SX960
5 -- 2 -- 4 -- B5 -- 75 MHz -- 50 MHz -- Q0704, Q0666, SX975, SX961, SZ977 (*) Not affected!
5 -- 2 -- 4 -- B5 -- 90 MHz -- 60 MHz -- Q0653, Q0654, Q0655, SX957, SX958, SX959, SX978 (*) Not affected!
5 -- 2 -- 4 -- B5 -- 100 MHz -- 66 MHz -- Q0656, Q0657, Q0658, SX962 (*) Not affected!
5 -- 2 -- 5 -- C2 -- 75 MHz -- 50 MHz -- Q0700, Q0725, Q0749, SK079 (*) Not affected!
5 -- 2 -- 5 -- C2 -- 90 MHz -- 60 MHz -- Q0668, SX999, SZ995, SU031 (*) Not affected!
5 -- 2 -- 5 -- C2 -- 100 MHz -- 50/66 MHz -- Q0697, Q0698, SX963, SX970, SZ996, SU032 (*) Not affected!
5 -- 2 -- 6 -- E0 -- 75 MHz -- 50 MHz -- Q0837, Q0846, SY005, SY009, SU097, SU098 (*) Not affected!
5 -- 2 -- 6 -- E0 -- 90 MHz -- 60 MHz -- Q0783, SY006, SL2WW, (*) Not affected!
5 -- 2 -- 6 -- E0 -- 100 MHz -- 50/66 MHz -- Q0784, SY007, SU099, SU110 (*) Not affected!

Pentium Overdrive for 586 systems --> Socket 5/7
Product Code -- Type -- Family -- Model -- Stepping -- Core stepping -- Clock rate -- FSB speed -- sSpec -- Version
PODP3V125 -- 0 -- 5 -- 2 -- cC -- aC0 -- 125 -- 50 -- SU081 -- 1.0 (*) Not affected!
PODP3V150 -- 0 -- 5 -- 2 -- cC -- aC0 -- 150 -- 60 -- SU083 -- 1.0 (*) Not affected!
PODP3V166 -- 0 -- 5 -- 2 -- cC -- aC0 -- 166 -- 66 -- SU084 -- 1.0 (*) Not affected!
PODPMT60X150 -- 1 -- 5 -- 4 -- 4 -- oxA3 -- 125/150 -- 50/60 -- SL24V -- 1.0 (*) Not affected!
PODPMT66X166 -- 1 -- 5 -- 4 -- 4 -- oxA3 -- 166 -- 66 -- SL24W -- 1.0 (*) Not affected!

Note: The attached picture is a Pentium 66 processor based on SX837 core stepping, which has the infamous FDIV Pentium bug, I have destroyed it a long time ago being useless for me! [Opened back cover and removed all pins.]

Last edited by karakarga on 2023-02-17, 21:49. Edited 67 times in total.

Reply 1 of 8, by H3nrik V!

User metadata
Rank l33t
Rank
l33t

Since 120 exists both in 600nm and 350nm, isn't it possible that some may have the bug as well, or has that been confirmed is not the case?

If it's dual it's kind of cool ... 😎

--- GA586DX --- P2B-DS --- BP6 ---

Please use the "quote" option if asking questions to what I write - it will really up the chances of me noticing 😀

Reply 2 of 8, by karakarga

User metadata
Rank Newbie
Rank
Newbie

Starting from 120 MHz and up, the bug was corrected on all versions!

Reply 3 of 8, by Deunan

User metadata
Rank l33t
Rank
l33t

This was some time ago but I did some minor research on the subject of Pentium FDIV bug - I think the conclusion was that anything Socket 5 was not affected.

I was mostly interested in obtaining a sample with broken FDIV but I didn't want to buy a ton of CPUs just to figure they are OK after all, so perhaps there are some faulty Socket 5 chips but very rare, that's why I decided to not bother with Socket 5. Frankly though I no longer remember the exact details, just the general conclusion I've made back then.

Reply 4 of 8, by AlessandroB

User metadata
Rank Oldbie
Rank
Oldbie

how to do the test??

Reply 5 of 8, by karakarga

User metadata
Rank Newbie
Rank
Newbie

There are various ways, watch https://www.youtube.com/watch?v=2OgP2WH5l_0 so paste "(4195835 / 3145727) * 3145727 - 4195835" to calculator. If result is 0 (zero) it is fine, if says -256, than it is a buggy fpu based chip.

So what does it mean? If you are a C/C++, Fortran, Pascal etc. languages programmer, the compiled programme might not work properly, especially when applying <include.math.h> in C/C++ language programme! So, it is adviceable not to use those buggy processors, when working on programming not to waste your time....

Reply 6 of 8, by skyfoxxp

User metadata
Rank Newbie
Rank
Newbie
Deunan wrote on 2023-01-18, 11:27:

This was some time ago but I did some minor research on the subject of Pentium FDIV bug - I think the conclusion was that anything Socket 5 was not affected.

I was mostly interested in obtaining a sample with broken FDIV but I didn't want to buy a ton of CPUs just to figure they are OK after all, so perhaps there are some faulty Socket 5 chips but very rare, that's why I decided to not bother with Socket 5. Frankly though I no longer remember the exact details, just the general conclusion I've made back then.

Back in the days in late 1994, my Gateway 2000 P5-90 bought in June 1994 was affected by the FDIV bug, and Intel offered me a free replacement, which I did.
So yes, some socket 5 CPUs are affected !

Reply 7 of 8, by shamino

User metadata
Rank l33t
Rank
l33t
skyfoxxp wrote on 2025-01-19, 16:50:
Deunan wrote on 2023-01-18, 11:27:

This was some time ago but I did some minor research on the subject of Pentium FDIV bug - I think the conclusion was that anything Socket 5 was not affected.

I was mostly interested in obtaining a sample with broken FDIV but I didn't want to buy a ton of CPUs just to figure they are OK after all, so perhaps there are some faulty Socket 5 chips but very rare, that's why I decided to not bother with Socket 5. Frankly though I no longer remember the exact details, just the general conclusion I've made back then.

Back in the days in late 1994, my Gateway 2000 P5-90 bought in June 1994 was affected by the FDIV bug, and Intel offered me a free replacement, which I did.
So yes, some socket 5 CPUs are affected !

The best thing about that is the replacement was probably more overclockable than the older stepping. 😀

Reply 8 of 8, by skyfoxxp

User metadata
Rank Newbie
Rank
Newbie
shamino wrote on 2025-01-20, 09:52:
skyfoxxp wrote on 2025-01-19, 16:50:
Deunan wrote on 2023-01-18, 11:27:

This was some time ago but I did some minor research on the subject of Pentium FDIV bug - I think the conclusion was that anything Socket 5 was not affected.

I was mostly interested in obtaining a sample with broken FDIV but I didn't want to buy a ton of CPUs just to figure they are OK after all, so perhaps there are some faulty Socket 5 chips but very rare, that's why I decided to not bother with Socket 5. Frankly though I no longer remember the exact details, just the general conclusion I've made back then.

Back in the days in late 1994, my Gateway 2000 P5-90 bought in June 1994 was affected by the FDIV bug, and Intel offered me a free replacement, which I did.
So yes, some socket 5 CPUs are affected !

The best thing about that is the replacement was probably more overclockable than the older stepping. 😀

Yes maybe, however the Intel Plato only had a jumper to set frequency to 75 / 90 / Reserved. "Reserved" was actually 100 MHz. 😉