So how many cores are enough these days?

[Eudimorphodon]

Oh no, totally with you there, but I nonetheless find it sort of depressing. IBM probably wasn't even unusually bad for the time; HP's various "small computer" product lines were at least as all over the map. And then of course you had DEC selling both PDP-8 and PDP-11 retreads alongside Z-80s and 8088s (sometimes stuffed into the same case)...

 

[Scali]

Oh no, totally with you there, but I nonetheless find it sort of depressing. IBM probably wasn't even unusually bad for the time; HP's various "small computer" product lines were at least as all over the map. And then of course you had DEC selling both PDP-8 and PDP-11 retreads alongside Z-80s and 8088s (sometimes stuffed into the same case)...

Yea, if you look at Commodore... Arguably the most successful computer company of the 80s, with the VIC-20, C64 and the Amiga... But they also had some nasty failures as well. C16, Plus/4, and the schizophrenic C128 I already mentioned... Weird designs, not fully taking advantage of what others in the company were doing.

 

[Agent Orange]

At one small moment in time yes
This forum goes a bit further back than most
Prior to the K7 architecture, AMD never threatened Intel's faster CPUs, and was merely price-fighting the budget models. Which in those days generally meant last-gen CPUs. Eg, when AMD had its first successes with the 386DX-40, Intel was already offering a 486DX-50.
AMD hit the sweet-spot for users on a budget (it was very popular among students for example).

The vendors cost for the Intel 486DX/50 back in 1992 was @ 570$ in quantities of 1000. The AMD AM386/40 was priced @ $40 more than the Intel 80386/33, or @ $278. At that period time, the AM386/40 rivaled the performance of Intel's 486/50DX with the exception of the on-board cache and FPU. The AM386/40 was one of the most successful CPU's of all time in terms of performance and sales.

 

[Caluser2000]

Also from memory the 486DX50 sales lost out to the 486DX2/66 of both Intel and AMD anyway.

 

[Scali]

At that period time, the AM386/40 rivaled the performance of Intel's 486/50DX with the exception of the on-board cache and FPU.

Erm lol!?
Yes, if you discount what makes a 486 a 486 (you forgot localbus), yes, then a 386 is almost as good.
But guess what, all 486es did come with on-board cache and an FPU, and that did make them considerably faster than a 386DX at the same clockspeed.
In fact, games generally did not use the FPU, and even then, a 486SX-25 was about equal to a 386DX-40 in most games.
See here for example: http://business.highbeam.com/409748/article-1G1-13042739/486sx25s-vs-386dx40s-upstart-fights-back

The tables turned when we put the machines through our performance tests. The 486SXs recorded clearly superior times on our application-based benchmarks: On average, the Intel-based systems were 10 percent faster than the AMD-based ones.

All five of the fastest systems in the review are 486SXs.

A 486DX-50 was a much faster system, especially when coupled with a localbus multi-IO card and videocard.

AM386/40 was one of the most successful CPU's of all time in terms of performance and sales.

Have any numbers to back that up?
Because back in the day I saw more 486DX2-66 systems than 386DX-40 ones.
While the 386DX-40 certainly was among the more successful CPUs, I highly doubt it's one of the most successful in terms of sales.
As for performance, see above.

 

[Scali]

Also from memory the 486DX50 sales lost out to the 486DX2/66 of both Intel and AMD anyway.

Well yes, the DX50 was quickly replaced with the DX2-66, because the 50 MHz bus proved to be unreliable.
But the DX2-66 was introduced shortly after the 386DX-40, and I merely mentioned the DX50 because it was the fastest Intel available at introduction of the 386DX-40, to show how far apart AMD and Intel were at the time.
For most of its life, the 38DX-40 was indeed fighting the 486DX2 CPUs.

 

[Caluser2000]

For most of its life, the 38DX-40 was indeed fighting the 486DX2 CPUs.

And here I was thinking it was Intel 486SX class cpus.

 

[Scali]

And here I was thinking it was Intel 486SX class cpus.

I mean in terms of performance, obviously. The price argument is useless and boring. Price is dictated by performance. Of course slower CPUs are cheaper. How else would they sell them?
I'm an engineer, tech talk only please. Obvious price strategies are obvious, and need not be discussed. As I already said above, the Am386DX-40 hit a certain sweet-spot in terms of price/performance.

 

[Caluser2000]

I mean in terms of performance, obviously. The price argument is useless and boring. Price is dictated by performance. Of course slower CPUs are cheaper. How else would they sell them?
I'm an engineer, tech talk only please. Obvious price strategies are obvious, and need not be discussed.

C'mon it's obvious your fanboyism is showing. Intel could've dropped their prices as well they had the capacity and profit margin to do so.

 

[Scali]

C'mon it's obvious your fanboyism is showing.

I AM NOT AN INTEL FANBOY!
I HATE INTEL FOR KILLING OFF MOTOROLA AND GETTING US STUCK WITH x86 FOR SO LONG!
ARE WE CLEAR?

Also, if there's any fanboy we should be picking on, it's Agent Orange... Trying to compare a 386DX-40 with a 486DX-50 performance-wise. What a joke.

Intel could've dropped their prices as well they had the capacity and profit margin to do so.

They could, no doubt. But why would they? People were willing to pay a premium for more performance and/or the "Intel Inside" on their system. Maximizing profit is not the same as maximizing sales.
In that sense, nothing changed between then and now. AMD is still offering CPUs that only compete with Intel's low-end/previous gen stuff, and AMD still has to undercut Intel prices because their brand name isn't that strong.

 

[krebizfan]

The AMD 386-40 was the typical AMD success story: a faster version of last generation's chip design when Intel was having yield problems with the new generation (keeping prices excessively high) and Intel stopped development of the older chip (to prevent Intel from cannibalizing Intel sales). The fast 286, fast 386, the various 486 and Pentium competitors all kept AMD in the game since lower R&D costs mean profits at lower prices. The Pentium 4 versus Athlon period was one where AMD lucked out. While performance generally tended to go in Intel's favor, the Pentium 4 used so much more power that AMD was able to sell at higher prices than normal. Now, AMD is behind on performance per core and power consumption which leaves price slashing as their only marketing method.

One fan page for 386-40 http://redhill.net.au/c/c-4.html

 

[Scali]

the Pentium 4 used so much more power that AMD was able to sell at higher prices than normal.

Another problem was Intel's RAMBUS deal: Intel was not allowed to offer DDR as an option, so you could only choose either an underperforming SDR system, or an overpriced RAMBUS system.
Intel actually already had DDR in the i845 chipset, which was used for SDR at first. They just couldn't produce DDR boards until the RAMBUS deal ran out.

AMD could use DDR from the start, so they could distance themselves from the Pentium 3, and the cheaper Pentium 4s with SDR were an easy prey. The RAMBUS machines were faster, but much more expensive.

 

[Chuck(G)]

Oh no, totally with you there, but I nonetheless find it sort of depressing. IBM probably wasn't even unusually bad for the time; HP's various "small computer" product lines were at least as all over the map. And then of course you had DEC selling both PDP-8 and PDP-11 retreads alongside Z-80s and 8088s (sometimes stuffed into the same case)...

I recall attending a CYBER 180 technical briefing in the mid 1970s at CDC. About halfway through the presentation, my patience finally got the better of me and I asked if the presenter was familiar with the STAR (a 1969 64-bit virtual-memory vector super). He said that no, he wasn't. I shot back that it was incredible that he knew nothing of an ongoing project that was located just three doors down the hall from him. I invited him to drop by to learn what lessons we learned, instead of re-inventing the wheel. Hey, it was morning and I hadn't had my morning cuppa...

To be fair, I knew nothing of CDC's military minis at the time.

 

[Caluser2000]

I AM NOT AN INTEL FANBOY!
I HATE INTEL FOR KILLING OFF MOTOROLA AND GETTING US STUCK WITH x86 FOR SO LONG!
ARE WE CLEAR?

You mean they lost out to competition? Diddums. Btw there's summit wrong with your keyboard.

 

[Scali]

You mean they lost out to competition? Diddums.

What's your point?
Besides, I wouldn't call it 'competition' as much as brute force. Intel's x86 just happened to become the choice for the IBM PC platform, which just happened to become by far the most popular platform ever. This allowed Intel to sink endless amounts of money into R&D and new fabs.
At least AMD is lifting along on the x86/PC popularity. All other CPU architectures are not, and are pretty much without a chance.
ARM being the exception, because it just happened to become the 'x86' of the mobile/embedded market... although it remains to be seen how long they will last now that Intel is targeting them.

Btw there's summit wrong with your keyboard.

No, there's something wrong with your eyes, because I already explained this earlier, and apparently it didn't register with you the previous time: http://www.vintage-computer.com/vcf...res-are-enough-these-days&p=355338#post355338

 

[Caluser2000]

I was actually refering to your caps lock but thank you for pointing out your older post. Hater, sorry disliker, of AMD fanboyz you say hmmm.

Yip how Intel approach ARM arch will be interesting. It's been around for a long time.

For the record I didn't quite get AOs comparison between the 386DX40 or the 468DX50 either.

 

[Scali]

Hater, sorry disliker, of AMD fanboyz you say hmmm.

All that "rooting for the underdog" gets old... Been hearing the same nonsense for over 20 years.

For the record I didn't quite get AOs comparison between the 386DX40 or the 468DX50 either.

Yea, my point exactly. It was annoying back then, and it still is just as annoying today.
Sure fine, the AMD 386DX40 was a good product, and offered great value for money. I said that myself. But let's end it there.
Trying to make outrageous claims about how technically advanced it is (it's literally a copy of Intel's 386, and since Intel was selling 486es at 50 MHz, surely they could have easily made 386es at 40 MHz as well, if it wasn't for the fact that they simply didn't have any room for that in their product lineup), or about what a great performer it was... now that's just fanboyism.
I got tired of dealing with that drivel many years ago. I certainly did NOT come here to re-live THAT particular part of my computer-related past.

Then again, I suppose I am a complex individual... On the one hand I don't have any love whatsoever for x86, seeing as I was basically forced to go from programming assembly on the sophisticated 68k and PPC architectures to the crude and archaic x86 architecture. On the other hand... in the fact that x86, MS-DOS and the IBM graphics standards are so primitive compared to their contemporaries, lies a big and interesting challenge as a programmer. So in a way I do enjoy coding routines for PCs, if only to make them do things that are easy to do on C64, Amiga and such, and deemed impossible on PC. Sometimes, suckage is a good thing

 

[Eudimorphodon]

So whatever happened to Cyrix, anyway? The 6x86L was pretty darn hawt for the dollar back in the day.

 

[Agent Orange]

Erm lol!?
Yes, if you discount what makes a 486 a 486 (you forgot localbus), yes, then a 386 is almost as good.
But guess what, all 486es did come with on-board cache and an FPU, and that did make them considerably faster than a 386DX at the same clockspeed.
In fact, games generally did not use the FPU, and even then, a 486SX-25 was about equal to a 386DX-40 in most games.
See here for example: http://business.highbeam.com/409748/article-1G1-13042739/486sx25s-vs-386dx40s-upstart-fights-back

A 486DX-50 was a much faster system, especially when coupled with a localbus multi-IO card and videocard.



Have any numbers to back that up?
Because back in the day I saw more 486DX2-66 systems than 386DX-40 ones.
While the 386DX-40 certainly was among the more successful CPUs, I highly doubt it's one of the most successful in terms of sales.
As for performance, see above.

I said "one" of the most successful not "the' most. I don't know where you where geographically back in the early 90's, but the AM386/40's were everywhere on this side of the pond. It would have been real hardship and almost cost prohibitive for the up and coming enthusiast to afford a 486 in the early going. Also, no one is disputing the prowess of the Intel 486 - it just wasn't within reach at that time. BTW - I presently have a AM386/40 with a co-processor up and running WIN 95. FWIW, I like it a lot better than my first 486SX, which was a real 'dog' in my opinion.

 

[Scali]

So whatever happened to Cyrix, anyway? The 6x86L was pretty darn hawt for the dollar back in the day.

They didn't do too well financially, and got bought by VIA in 1999. As far as I know, VIA still makes low-power x86-compatible CPUs today, mainly for the embedded market.