What retro activity did you get up to today?

Discussion in 'Retro & Arcade' started by adz, Jan 28, 2014.

  1. philscomputerlab

    philscomputerlab Member

    Joined:
    Sep 29, 2014
    Messages:
    1,843
    Look Shane, a joke is a joke, but you're dragging it along for too long now and then a joke turns into hassling others.

    Every time someone mentions Pentium 4 you got a stupid comment to make and it's really getting annoying. You might think it's funny, but I'm calling you out as a bully.

    People should share their retro activities without getting picked on all the time. I don't know what your angle is, but I'd like you to stop please.
     
  2. qwertylesh

    qwertylesh Member

    Joined:
    Aug 21, 2007
    Messages:
    8,931
    P4 hate is rather valid though phil.

    It's lifecycle was long and there were a lot of issues with many steppings released.

    people who rag on P4 usually had bad experiences with the 478 variances, where they ran abnormally hot (enough so that just windows loading up would cause the fans to go bonkers)

    I take it you didnt experience it?


    I dont see shane dishing out any more then he gets, but if he's getting too much for you I can cryovac him in my foodsaver if you like. :cool:
     
  3. philscomputerlab

    philscomputerlab Member

    Joined:
    Sep 29, 2014
    Messages:
    1,843
    No, I'm not experiencing this I'm afraid. Neither did I when I had a P4 back in the days.

    Sure maybe it's because I'm not using the stock cooler, but don't we all use after-market coolers? Heat shouldn't be a problem for socket 478. For processor options it tops out at 3.4 GHz and you have a range of cores to choose from: Northwood, Prescott and Gallatin.

    Later processors for 775, now that's another story.

    Now you mention issues, and I couldn't disagree more, with the P4 you really have an issue free retro life.

    I found lots of little challenges working with Athlon XP machines. I do plan to make a video about this at some stage.
     
  4. elvis

    elvis Old school old fool

    Joined:
    Jun 27, 2001
    Messages:
    43,983
    Location:
    Brisbane
    I can't speak for others, but I'll speak my mind on the matter.

    The P4 was a terrible architecture. It represented a time in Intel's lifespan when, for the first time in the company's long history, their CEO was a person from a business/marketing background, and not a technical background.

    What resulted was the P4 netburst architecture - a thinking that it didn't matter what the market actually wanted, but rather what Intel could milk out of a poor design.

    Netburst has shocking FPU performance, which resulted in shocking 3D performance. What they did find however was that they could ramp the clockspeed of these things up to silly numbers. That sounded like a good idea when it came to marketing, but people quickly learned that insane clockspeeds were worthless when RAM was slow at the time, and the resulting differences between FSB clocks and the CPU clocks meant that these "super fast" CPUs sat around doing not much, starved by a lack of information fed to them.

    Couple this with the fact that the 3D market was starting to boom. Both professional 3D users (CAD/CAM/DCC/VFX) and consumer 3D users (gamers) wanted solid floating point performance, and Netburst did little to satisfy that need.

    AMD knew they had a better designed product. They knew that their K6 architecture was weak at floating point, and with a recent acquisition of DEC Alpha engineers, had invested a lot of time and money into ensuring their upcoming K7 (Althlon) architecture would meet that emerging market's need. The "Thunderbird" Athlons with their incredible floating point power beat the pants off Intel P4 chips early in the race.

    What they found, however, is that they couldn't match the clockspeeds of Netburst over time, as Intel found that they could keep jacking up the P4 clockspeed easily, even if it didn't really help performance linearly. Sadly with a very uneducated public, they needed a way to combat this. So the "XP +" naming convention came about. The Althon XP 1500+ only ran at 1333MHz, but easily matched a 1.5GHz P4 for performance.

    This went on for a while, and worked pretty well for AMD.

    Eventually Intel realised what a terrible architecture Netburst/P4 was, and also what a terrible CEO they had. They sacked the marketing dick, and replaced him with an engineer. The first thing that happened after that was Intel started looking at the mobile market. R&D into the older PIII architecture showed that they could get much better performance clock-for-clock, and especially per watt out of the old architecture with some tweaks. They did that, and made the Pentium M. More R&D then went on to produce the "Core" processors (again, historically linked to the PIII and not the P4). On and on Intel go, to what we have today in the i7 series, which are a direct descendent.

    I'm particularly thankful for AMD's existence during this period. They might be struggling today, but without them we'd be stuck on a really shitty platform in x86 land. The same can be said for 64bit - Intel's plan was to move everyone to Itanium (IA64), which is a terrible platform (I've worked on it professionally on HP-UX servers, and it sucked). AMD extended x86 to x86_64 (the correct name for which is actually AMD64, which confuses people who run Intel chips that are the AMD64 architecture, but quite correctly so the patent belongs to AMD).

    AMD showed the world that competition is ALWAYS good, and without it the market leader WILL get lazy, and WILL rip off their customers. Good engineering matters, and tech companies at the top tend to forget this (all the big guys have done it - Microsoft and Apple in the software world are two that I consider very sloppy right now, and in need of a huge kick up the arse just like Intel got in the P4 days).

    While plenty of folks on these forums are probably too young to remember it, I lived through all that professionally, having to deal with P4/Netburst architecture while running the IT for a large international architecture firm that specialised in both CAD and 3D renders of sporting stadia around the world (we did the Suncorp Stadium, Stadium Australia / Telstra Stadium / whatever it's called now for the 2000 Olympics, Wembly Stadium in the UK, and a tonne of NFL and Baseball stadia throughout the US). Grunty 3D was mandatory for us back then, and the P4 let us down time and time again. Where we could, we turned to AMD. But that was often quite difficult when dealing with name-brand corporate vendors like Dell and HP.

    So a tip of my hat to AMD. Their Athlon family gave the world a great processor while Netburst floundered, with it's enormous power usage, crazy heat problems, and woeful FSB to core clock ratios that starved an already weak CPU. It also forced Intel to sack their shitty CEO, and go back to their roots of being a good technical company lead by good technical people, rather than yet another marketing-driven bunch of suits.

    And that's why I don't like the P4, and would rather play with old Athlons any day of the week. FWIW, all my MAME arcade machines run AthlonXP processors, and are still running great today (they've never needed to be upgraded). That's also part of the reason why I struggle calling them "retro", because for me they're still in use as they were when I installed them.
     
  5. philscomputerlab

    philscomputerlab Member

    Joined:
    Sep 29, 2014
    Messages:
    1,843
    Thank you for explaining your view, very interesting!
     
  6. shane41

    shane41 Member

    Joined:
    Nov 10, 2008
    Messages:
    6,936
    Location:
    dice que no sabe
    Edit Just notice elvis hit with a comprehensive reply.

    As querty said, he nailed it.

    Look I have plenty of 478 boards here....... they'll just sit in a cardboard box somewhere. Certainly not excited to post in retro about them. There were plenty of them made & I'm still finding them pop up all over the place. Like a damn virus.

    Occasionally someone gives me a pc case & I'll find a slot board in it. That's a nice surprise. :) But that's a very rare surprise now, usually 98% P4.

    You can post all the 478 stuff you want Phil, I'm not holding you back.
    Just a stir m8, you all too serious.

    Hope / wish the guys strive to find / build better machines that's all. Still plenty of choices out there, even Atmo posted a decent 370 board $25 on ebay. Lift the bar chaps :thumbup: P4 is too lazy / easy
     
  7. demiurge3141

    demiurge3141 Member

    Joined:
    Aug 19, 2005
    Messages:
    2,236
    Location:
    Melbourne 3073
    Though rare these days, the best AGP boards are socket 479 ones that runs mobile banias/dothan pentium-m's. I think AOPEN and DFI had a few models.
     
  8. Cannula

    Cannula Member

    Joined:
    Jun 9, 2002
    Messages:
    1,749
    Location:
    Lisarow 2250
    Don't forget the awesome Asus boards that use the CT-479 adapter to turn a P4 board into a Pentium-M board.
    I'm sure Shane would give exception to those. ;)
     
  9. qwertylesh

    qwertylesh Member

    Joined:
    Aug 21, 2007
    Messages:
    8,931
    Hell the architecture isnt even why i'm personally not fond of that era of intel tech.

    the bad cap plague from that era :Paranoid:
     
  10. shane41

    shane41 Member

    Joined:
    Nov 10, 2008
    Messages:
    6,936
    Location:
    dice que no sabe
    Actually I was going to post you'd be along :lol:

    Quert Just too complicated to post all that stuff elvis did.
    50% of it is relevant to my thoughts & knowledge at that time.
    But explaining it makes my brain hurt, I'd rather go & link it up. :D
     
  11. elvis

    elvis Old school old fool

    Joined:
    Jun 27, 2001
    Messages:
    43,983
    Location:
    Brisbane
    Meh, capacitors are cheap. Buy a few dozen rolls and fix yourself some busted boards.

    I rescued quite a few motherboards and video cards back in the day by replacing bad caps. You don't even have to be a very competent solderer or understand electronics. Match the capacitance ratings, ensure the voltage ratings are the same or a little higher, and match the polarity (some photos before you start can save your hide).

    If you call yourself a retro enthusiast, but you haven't taken a soldering iron to a motherboard yet, then you've got some work to do to earn your stripes. :)
     
  12. DonutKing

    DonutKing Member

    Joined:
    Mar 21, 2004
    Messages:
    1,426
    Location:
    Tweed/Gold Coast
    In my experience, at the time of the P4's introduction - Athlon really was the much better product, and cheaper too.

    Intel was pushing RDRAM which was hideously expensive at the time. The alternative was SDRAM which really choked the Netburst architecture; when paired with a Willamette it made for a bog slow machine. I knew a bloke who paid up for a P4 when it first came out but got SDRAM with it; the machine felt amazingly sluggish, more so than a P3 of about 2/3 the clock speed.

    So if you paid the premium for a P4 plus RDRAM and a mobo to suit, you'd probably be out several hundred, if not a grand, compared to an AMD system with SDRAM that could generally keep up with the P4 except for some edge cases (mainly synthetic benchmarks that supported SSE2 like Sisoft Sandra)

    Even the P3 was a faster CPU than the Willamette P4 clock for clock (except for SSE2 optimized applications). Of course, the P4 could scale to 2GHz and beyond which the P3 couldn't... plus, Intel jacked up the price of high end P3's so as not to cannibalize their P4 sales.

    Basically Intel's shitty marketing and dodgy deal with Rambus got a lot of people's backs up and pushed people over to AMD.

    In 2002 Intel bought out the Northwood core and DDR chipsets which really evened up the playing fields, however AMD still had a very competitive product with the Athlon XP and Athlon 64 for quite a while yet. Intel was still generally more expensive for a comparable system.

    Yes the Athlon's ran hot and had fragile cores, but the Willamette's ran pretty hot too... and I never crushed a core myself, despite swapping CPU's and heatsinks several times. :)

    As for dodgy motherboards and chipsets... I've seen dodgy boards on both sides, and not just the ubiquitous bad caps. Then I had an ABit NF7 which lasted years until I sold it.

    I was nearly going to post something about how this hardware isn't really 'retro' but then I realised I enjoyed reminiscing about them; even if I wouldn't bother tracking down or trying to build such a system. :)
     
  13. elvis

    elvis Old school old fool

    Joined:
    Jun 27, 2001
    Messages:
    43,983
    Location:
    Brisbane
    Ah yes, I forgot about that. Rambus' design was proprietary, and no other manufacturer could make it (unlike SDRAM, which is made by many different vendors).

    And what happens when something is proprietary and closed? There's no competition, and it becomes overpriced. It always amazes me when self-proclaimed capitalists attempt to thwart the natural flow of the free market, and wonder why their paying customers get pissed off and go elsewhere.

    Yes, the P4 era was a dark time indeed for x86. I'm so very glad it's all over now, and Intel learned from their mistakes. Once again, thank the good lord for competition, and in this specific case AMD. While AMD may have fallen out of favour with high end PC enthusiasts today, their efforts in quality R&D during those days will always be appreciated.

    Yup. And again, once Intel got rid of their idiot "marketing-guy" CEO, they went back to that superior P3 architecture for the Pentium-M, Core, and onwards.

    The P4 was a very wrong turn for Intel. Backtracking to the P3 and taking it forwards was a good strategy from both technology and business points of view.
     
    Last edited: Mar 26, 2016
  14. shane41

    shane41 Member

    Joined:
    Nov 10, 2008
    Messages:
    6,936
    Location:
    dice que no sabe
    Curious Phil, do you have a slotA board there in your arsenal?
     
  15. mAJORD

    mAJORD Member

    Joined:
    Jun 4, 2002
    Messages:
    11,817
    Location:
    Griffin , Brisbane
    One thing that's of often not mentioned though is that Whilst Netburst was pretty horrible overall (though it had its ups and downs), the architecture had a lot of very novel tricks up its sleeve for its time, to try and combat is ridiculously long pipeline.. some of which have eventually returned inside more modern architectures. e.g:

    - Sandy bridge's uOP Cache, whilst different in purpose is the same concept to Netburst's Trace cache.

    - AMD's bobcat , Bulldozer and Sandy bridge all moved to using Physical register files for register renaming, again first seen on netburst.

    It's the one good thing that can come out of ridiculous management decisions and expectations- they put up an engineering challenge that forces innovation
     
  16. qwertylesh

    qwertylesh Member

    Joined:
    Aug 21, 2007
    Messages:
    8,931
    I've done hundreds of recaps actually, mostly network hardware though mobos and gpus don't seem to be any more difficult and yes its very easy to do.

    I just didn't do mobo recaps when I worked in IT retail, it wasn't worth the time or effort to keep customers on out of date hardware.

    Well sure, the same can probably be said for nVidia's fermi (Geforce 4xx series), when you're comparing the feature sets of the newer and better breeds of hardware, but that doesn't excuse or void the fact that fermi was hot unstable totally bullshit rubbish hardware that should have never made it to market.

    or hell, the hdd killer that is Vista could be put under the same analogy, but i still would have preferred if its benefits made it to newer products while it never seeing the light of day.

    anyway, I hate both above examples way more then P4, I owned a rather high qty of P4's in my time, and i have had systems from both camps in that era, both run reliably and instances where they've ran terribly.

    personally, if i were to get into retro pc, i'd be goin for 486dx100 98se, S1/A PIII 500 2k sp4, and probably a qx9650 xp machine, but that's me.
     
    Last edited: Mar 27, 2016
  17. shane41

    shane41 Member

    Joined:
    Nov 10, 2008
    Messages:
    6,936
    Location:
    dice que no sabe
  18. gdjacobs

    gdjacobs Member

    Joined:
    Apr 3, 2007
    Messages:
    2,637
    Location:
    MB, Canada
    Intel's prefetch and speculative execution capabilities were very advanced by the end of the P4 era. Again, necessary to combat the high latency and long pipeline of the P4 platform.
     
  19. elvis

    elvis Old school old fool

    Joined:
    Jun 27, 2001
    Messages:
    43,983
    Location:
    Brisbane
    Yup, that's fair enough. Particularly if you've got the manufacturer's RMA processes there to cover it. That's their job.

    Obviously for our hobby purposes, that's a different thing all together. Often a complete recap will keep a board in top working order, and improve stability no end. I know a lot of guys who do the same with CRT chassis - if they buy a second hand one, they'll immediately do a recap before bothering to power the thing up, as they know that they can buy better quality caps today than either the original manufacturer would have used, or any third party could have used on repairs years ago.
     
  20. qwertylesh

    qwertylesh Member

    Joined:
    Aug 21, 2007
    Messages:
    8,931
    I actually want to collect way older systems.

    I have a CPC464, and I've wanted to get a C64 bread-bin for the longest time, also a zx spectrum
     

Share This Page

Advertisement: