[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr / vt ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

Due to resource constraints, /g/ and /tg/ will no longer be archived or available. Other archivers continue to archive these boards.Become a Patron!

/g/ - Technology

View post   

[ Toggle deleted replies ]
File: 54 KB, 832x760, intel-core-2-duo-e8400-3-0-ghz-6mb-1333-mhz-socket-775-original-imaefzr37z5jnng6.jpg [View same] [iqdb] [saucenao] [google] [report]
71135087 No.71135087 [Reply] [Original] [archived.moe] [rbt]

Zoomer here.
How did people react to the original multi core processors?
Were they immediately well received or seen as more of a gimmick?

>> No.71135094


>> No.71135163
File: 125 KB, 1000x1000, 1554170985946.png [View same] [iqdb] [saucenao] [google] [report]

>Holy crap Lois! It's like 2 CPUs for the price of 1!!!

>> No.71135192

They're still mostly a gimmick today. All the actually interesting shit cannot be parallelized well.

>> No.71135199

Had an Hp-Compaq with that, still works but gave it away because it overheats even with new paste and cleaning.

>> No.71135210

It was common for enthusiasts to run dual cpu setups so people were happy they could get two in one

>> No.71135218

AMD was revolutionary back then. First multi core, breaking the 1GHz barrier, first memory controller onboard, first 64bit processor
Jim keller is the man.

>> No.71135239

Back then I was a gaymer and in that community it was seen as a gimmick.
You should choose the right hardware for the job. The thing is, I don't even care if multithreading is a gimmick 99% of the time, I made an informed decision and bought my CPU for a very specific purpose. I am very happy with it.

>> No.71135244

i don't want to sound like a boomer or anything
but the lack of cows in this thread is shameful.

>> No.71135264

O remember the original jump to x64, my father was going crazy how he was getting double the power for cheep.

>> No.71135277

People did not know exactly what to do with them and thought it was complicsted

>> No.71135421
File: 140 KB, 1038x1000, 1523433903969.jpg [View same] [iqdb] [saucenao] [google] [report]

>memory controller onboard
That shit still amazes me to this day, Intel didn't have an on chip IMC until 2008 with Bloomfield i7 and AMDhad it in 2003.

>> No.71135433

Originally they were absolutely dominated by AMD, and Intel dual cores were for retarded fuckheads and shills. In a couple years Intel took the lead back. It was the golden age for AMD. Ryzen hopes to bring back that magic by dominating i9 but we will see. It won't be the same as before, even if it is better on paper.

>> No.71135449

Depends on who you asked. It was mostly a gimmick to gamers and other home users who were running a handful of single-threaded applications at once. In the high-end it was a great way to reduce power consumption and increase density in multi-processor systems.
AMD fanboys are the mactards of computer hardware. 64-bit and multi-core existed in high-end machines for years prior to AMD finally bringing them over to the consumer realm.

>> No.71135450


>> No.71135469

>multi-core existed in high-end machines for years
Of course espensive ass servers and dual processors existed prior to AMD bringing it to the masses, but the point is that AMD brought it to us for less money when Intel refused to innovate.

>> No.71135480


>> No.71135514

>coping this hard
There is a big difference between processors the with cooling the size of a fucking bean bag chair, and ones that you can fit in a laptop or a cheap desktop. Any retard can make the former, which is why so many retards were making them in the 80s.
Let's try not to pretend like Intel was dominant with the dual cores by downplaying AMDs achievements.

>> No.71135607

Who said anything about Intel? Early multi-core processors were mostly pioneered by IBM, HP and Sun, while credit for 64-bit goes to MIPS and DEC. I'm just reminding you that AMD did not literally invent modern computing just because they shit out a cheap dual-core chip at the same time everyone else was. Even Apple was going dual-core in 2005.

>> No.71135640

A lot of people said you wouldn't need a core2quad back in the XP days, iirc Windows 98 and 95 could only use 1 core or something.

>> No.71135641

They're innovators Anon, you don't get it, it's like how apple takes ideas and technologies that already exist, and package it in such away, you think "Why hasn't anyone done this?"

>> No.71135689
File: 10 KB, 326x383, 1294374058624.png [View same] [iqdb] [saucenao] [google] [report]

I remember when the q6600 came around and /g/ was flooded with cow2beef pictures

>mfw e8400fags got eternally btfo in the end

>> No.71135721

Single core fags will always be btfo, intel pretty much died tonight with the announcement of midrange 8 core that beat the 9900K and that 12 core part, wich I don't mind saying is a steal at $500 because intel wants $1100 for theirs.

>> No.71135785

Not really. Mine served me nearly a decade gaming. Worse performance but not in all games and not by much

>> No.71135835

Not sure what GPU you had, but a modern game like GTA V slams out a Q6600 and the bottlenecking is crazy even on a mid range GCN card, the game is not playable on wolfdales and that game came out in 2015.

>> No.71135861

Except they really didn't. Of the major desktop and server architectures that were still being actively developed at the time, x86 was the *last* architecture to get on the multi-core bandwagon. IBM was doing it since 2001. HP, Sun and even Itanium since 2004. 64-bit computing of course had already long been a thing since the early '90s, AMD and Intel only finally bothered to glue it on to x86 when Itanium flopped and consumer systems began to approach the 4 GiB memory barrier.

Of course, AMD does parallel with Apple in that they have done a great job of deluding their fanboys into believing they were revolutionaries for merely following market trends in a way that best pandered to their customer base, but at least Apple put their own original spin on the more complex ideas they took. You can't really do that with fundamental concepts like these.

>> No.71135867

That's a different Era in gaming and I actually did beat it on that cpu and 8800gtx on low. I had it on 4ghz though

>> No.71135891

How is life in eastern Slovakia/Bengladesh/Mexico/Philippines

>> No.71135907

>that cpu and 8800gtx
Damn i would've given my nuts for an E8400 rig with an 8800GTX, and a 4GHz wolfdale core is as good as bloomfield core without hyperthreading when it came to ipc, and it hit the magic 4GHz at the time.

>> No.71135944

I don't own it anymore. The video card fan went kaput and then the motherboard

>> No.71136151

How is packaging a 2nd core different than what i486 did with "upgradability"

Basically it is on one chip versus an expansion slot


>> No.71136189

>the absolute state of zoomers

>> No.71136194

Intel was pretty much always huge company that never innovated unless they had to.

>> No.71136219

Meanwhile most of AMD's "innovation" is from an ex-DEC engineer, kek. Fuck x86 kikes.

>> No.71136351

Most of the work done on most CPUs is small processes that do something and then exit. Multiple cores help tremendously with this. Very few computer users are running tasks which can’t be parallelized.

>> No.71136376

Intel bent the knee and hired the same fellow you mention because their Pajeets and Chaims just can’t innovate.

>> No.71136477

>smaller chip
>cheaper consumer oriented chip
>pretty much bought multicore and 64bit arch to masses

>not innovation
Before AMD, most people didn't heard any of this. IBM, sun, hp.... All failed
I know, you don't want to give any credit to AMD, but reality is that all these things AMD did brought to mainstream.

>> No.71137185

Technology is more than consumer garbage and gaming accessories where having the biggest marketing machine and selling the most == being the first, or the best.

AMD brought 64-bit and multi-core to the desktop PC mainstream by virtue of being the first of the big consumer technology juggernauts that bothered to care about it once it became clear their target market could benefit from it, just as the likes of MIPS and IBM did when they pioneered those technologies in the lower volume high-end years before.

AMD didn't succeed where others failed, they just took what they did successfully for their own users and sold them to a larger market in line with industry expectations. As I said, x86 was the last major desktop architecture to adopt multi-core and 64-bit, there is nothing innovative about that unless you restrict yourself squarely to the x86 bubble and ignore everything else outside of it, something /v/ denizens tend to do while ironically calling themselves tech savvy.

>> No.71137274

A gimmik but an impressive one. I remember talking about it with my buds and arguing whether it was better despite the lower.clockspeeds. Remember that this was after the p4 era when GHZ was everything.

Most people did not undertand what threads or cores were (nor what they implied in terms of real world impact). What most people believed was that a dual core was essentially 2 CPUs glued together.

By the time multicore processors were fairly common, the argument shifted to whether programs actually used those other cores, and whether it was better to buy a powerful dual core(E8400) as opposed to a entry level quad (Q6600)

>> No.71137370

As I already said, it seemed like a meme.
I understood more when I started programming.
If you are into tech learn to code a bit, you dont have to be a pro, a little bit is fun.

>> No.71137577

when these cpus were released, nearly nothing was using them. it took a while

>> No.71137627

I had my gentoo phase back in the day and Athlon X2 were seen as the holy grail.

>> No.71137854

Earliest dual core x86 I've seen were the double socket Pentium 3s. Most people who used those told they were pointless (remember, gaming OS at the time was Win98SE, which didn't even support dual cpus). On Win2k you could use both cpus, but gaming was limited, and the only way to use the extra cpu was running 2 apps at the same time, so this was useless for gaming.

First *mainstream* dual core cpu was the Athlon 64 x2, which was iirc 2 chips in a MCM module. Worked fine under WinXP, but very few games supported it. However, by that time you had things like FRAPS common, and you had Pentium 4s with Hyperthreading, so having two cores wasn't that rare and more and more things started supporting it.

I remember jumping from A64 to A64X2 in... 2005? 2006? I don't remember. The main advantages were faster video decoders for HD rips (at the time they were HDTV rips from MPEG2 TS files encoded into h264 or xvid), and some emulators supported it (PCSX2 straight up doubled in framerate and it was playable at full speed for the first time).

The biggest advantage for me was, when a dodgy app decided to hang up in an infinite loop, it did not make the entire OS unresponsive. The OS could run on the other core and I could alt-tab and nuke it in the task manager. This might seem insignificant, but when a fullscreen game did this on a single core machine, you had to either wait 15 minutes until it managed to open the task manager, or do a hardware reset. So that alone made it worth it for me, and I was playing a lot of PS2 emulated titles like Disgaea and Gradius V, and watching HD movies, so the Athlon64 X2 was fucking awesome at the time. For people playing GTA, FIFA and NFS, probably not so much.

Then the Core 2 increased performance a shit load over that, and the Q6600 double core count; AMDs response was buggy and slow and wasn't until a respin that it beat the Core 2 due to higher clocks and cores, but by that time Intel had i7s and i5s out.

>> No.71137870
File: 354 KB, 1356x1539, 1208553523125.png [View same] [iqdb] [saucenao] [google] [report]

Yeah, that was the time when this was made on /g/.

There's also a CUDA-optimized version the gives you a steak burned to coal in 0.2 seconds.

>> No.71137871

It's still a gimmick.

>> No.71137879

>First *mainstream* dual core cpu was the Athlon 64 x2, which was iirc 2 chips in a MCM module.
Athlon 64 x2 was monolithic. Pentium Ds were the MCM dual cores.

>> No.71137881

So why does AMD own the patents to 64bit tech?

>> No.71137885

>Technology is more than consumer garbage
Oh, sorry. I didn't know that we're talking from other perspective rather than consumer.
Tell me, about all these data centers and mainframes you own. :^)
Btw, stopped reading right there

>> No.71137997

They were? My bad then. All I remember is being stuck in Socket 939 during the DDR->DDR2 transition, then trying and failing to find an Opteron for it (the S939 Opterons were rebranded FX chips). Then when Vista came out, I had to swap out the motherboard in a hurry, because I used a S939 dual core with a Radeon card with an nForce3 chipset, and Nvidia DID NOT release Vista drivers for the nForce3. You could get it sort of stable with XP drivers, but that only worked if you used a Geforce card - with Radeons, it was unusable.
With a cheap VIA chipset board (remember those?) it worked fine even on Windows 7.

So basically, the most popular high-end PCs from 2003-7 just did not work under a new OS due to Nvidia being cunts. This fucked over not just me but most people I know who had high end PCs. I haven't bought nvidia cards ever since.

>> No.71138085
File: 188 KB, 720x1348, 1558899455041.jpg [View same] [iqdb] [saucenao] [google] [report]

AMD had better dual-core system from beginning. Intel had their D-furnaces with TDP above 90W.

>> No.71138188

They don't. They have patents on their specific means of extending x86 registers to 64 bits while maintaining compatibility with older 32-bit designs, but they did not invent 64-bit microprocessors or the multi-core concept. I don't get why this is difficult to understand.
>I'm ignorant so it doesn't count!
Just stop posting, you're a fucking idiot.

>> No.71138198


>> No.71138204

real brainlet hours

>> No.71138228

I don't have a picture if ur mum on my computer, sorry.

>> No.71138264

i didn't really pay it much attention cause i was busy fucking bitches. i did know of their existence though.

>> No.71138318

>somehow non consumer computers are more relevant now
How, exactly? Dumb ass nigger

>> No.71138334

>yfw still using the Q6600 today

Feels good man. Still runs everything I need it to. Only a retard needlessly upgrades often

>> No.71138335

Gimmick for sure.

We had hyperthreading and the only thing people cared about was single core performance for games.

So basically the same as now.

>> No.71138458

Did Apple invent the GUI and the Mouse because they were the first company to advertise it to average joes? How is this so difficult for you to comprehend? It doesn't matter.

>> No.71138519
File: 7 KB, 224x250, 1555647611661.jpg [View same] [iqdb] [saucenao] [google] [report]

>>somehow the computers that are actually doing the work to support my epic gamer toys and internet connected fashion accessories are more relevant now

>> No.71138522

Innovation and invention isn't same thing, dumb faggot
Innovation means introductory to something new
And AMD literally introduced multi core computing and 64bit arch to consumers.

>> No.71138536

They were received with the ultimate in retardation. Well, more accurately it was the single core CPUs with hyperthreading that were reported as if they were two separate CPUs, then the true dual core CPUs reporting the same. The other bit of retardation was how people would add the clock speed of cores together. Suddenly a 1GHz dual core CPU was the same as a 2GHz CPU to idiot consumers.

>> No.71138542

>How did people react to the original multi core processors?
The first two cores were an immediate improvement for general QoL, anyone that disagreed was doing so out of budget concern or ignorance, granted it was like 3-5 times the price for a CPU that wouldn't clock as high and need aftermarket cooling.
Quad cores when they first appeared were like 8+ cores now seen more as a niche product most people couldn't use properly than anything else, but most reviews still underlined that it would certainly end up being useful down the line and well Q6600 are mostly useable whereas dualcore Core 2 are really starting to become hard to use now

>> No.71138547

Fucking this. Exactly this. And the "2 cores, each running at 2GHz? I HAVE 4GHZ OF PROCESSING POWER NOOBS"

>> No.71138552

I didn't even mentioned application of computers.
And, obviously, implied In this place (/g/), dumb fuck.
Post pictures of your data centers or mainframes. Let me guess, you have none, but most definitely have a desktop or a laptop.
What a fucking faggot

>> No.71138564
File: 129 KB, 900x729, 1542186045917.png [View same] [iqdb] [saucenao] [google] [report]

>I didn't even mentioned application of computers.
lol. seriously just shut the fuck up retard.

>> No.71138573

I dunno, my dude. My first dual core CPU wasn't that impressive. I went from a P4 to a C2D, and the improvement was not nearly as drastic as you imply. Maybe it's because of the hyperthreading of the P4, but I was really underwhelmed considering how much I spent on my first C2D.

>> No.71138580

They were likened to dual processor systems. The Pentium D was Intel's first offering for consumers, basically two Pentium 4s on one core, and it was fucking horrible. A power hog that couldn't be cooled properly, sound familiar? The Core 2 Duo came along about a year later and was a HUGE improvement.

I believe AMD's first was the Athlon 64 x2 - pretty good chip, I used one for a few years and was perfectly happy with it, the Core 2 Duo was better though.

Back then you could get away with using a single core computer for everything. How things have changed.

>> No.71138642
File: 625 KB, 2752x1334, 1489181606975.jpg [View same] [iqdb] [saucenao] [google] [report]

Is the concept even still relevant to how multi-core and multi-threaded applications are handled these days?

>> No.71138643

>Innovation means introductory to something new
No, innovation just means something new. Multi-core and 64-bit were not new technologies in 2005, not even necessarily to consumers; see MIPS as an example of a 64-bit architecture widely used in low-cost embedded systems even before the days of the Athlon. Embedded is also another great place to find asymmetric multi-core designs as well, especially combined CPU/DSP chipsets like the OMAP that were of similar complexity.

>> No.71138656

Maybe it's you that should shut up? Not because you're wrong, but because this fairy farming cocktugger isn't going to stop being fucking retarded and it's only going to end when you let it go and realize you're right and continuing to argue is like winning the special Olympics

>> No.71138664

>implying programmer or a graphics designer, or an engineer using computer at home isn't consumer
You're retarded, aren't you?

>> No.71138674


>> No.71138698

> : the introduction of something new
> : a new idea, method, or device : novelty
Where's the reintroduction of something old to a new market, here? I don't see it.

>> No.71138731

new to the market senpai.

It's like saying 'X brought innovation to our office by implementing X technology!"

>> No.71138732
File: 87 KB, 971x565, 1553825752470.jpg [View same] [iqdb] [saucenao] [google] [report]

don't worry, I only ever saved like three cancer wojaks
>but what about the 2% of these professions that work from home
see pic also please learn english

>> No.71138755

Well my first experience with one was when the local cybercafe replaced one of it's dead computer (3500+) with a dual core capable box (4400X2)
Owner was running all sorts of spyware shit (including something that would basically reset the computer to it's original install at every reboot) so we couldn't do illegal shit on it's computers so the single core all choked like hell if you did much of anything whereas the single dualcore would still do just fine.
That and I was very much into ripping DVDs for my own to make the most of that DVD renting store so dualcore meant I could actually do shit while the movie was encoding, but I've always had an issue with my behavior leading top having an absolute clusterfuck of stuff open at all times so moar cores benefited me more than most I'd assume (Also why I'm most likely gonna pull the trigger on the 3900X once it's out for a little while just to avoid some nasty surprises)

>> No.71138757

Sounds more like mental gymnastics to me. Just accept that your faceless corporation of the month is not God.

>> No.71138776
File: 79 KB, 1099x605, 1540396635455.jpg [View same] [iqdb] [saucenao] [google] [report]

What you're even talking about? You moved goalposts beyond comprehension.

>> No.71138786

I'm not even the same guy ;_;
I have a Xeon E-2174G since the jews at dell didn't have a ryzen option.... I actually bluescreen once each day when doing heavy MEMORY ops... hold me

>> No.71138787

who are you quoting? learn english.

>> No.71138808

Based retard.

>> No.71138871

That sucks. EPYC doesn't seem to have gotten very far yet, but Intel probably won't cease being a circus of utter incompetence for quite a while so you'll hopefully have some better options by your next upgrade.

>> No.71138914

>Originally they were absolutely dominated by AMD, and Intel dual cores were for retarded fuckheads and shills.
Stop lying. AMD ruled the Pentium 4 era, just before the dual cores. Intel almost killed AMD with Conroe IPC and AMD played second fiddle until Ryzen.

>> No.71138917

Are you posting from 2014?

>> No.71138965
File: 30 KB, 550x278, amdx2_3800.jpg [View same] [iqdb] [saucenao] [google] [report]

>not posting the first dual core
truly a zoomer

>> No.71139071

that doesn’t look like POWER4

>> No.71139133

>Pentium D dun exist guyse

>> No.71139806

AMD dominated power wise, so Intel did what Intel does best, and bribed OEMs to only use Intel CPUs.

>> No.71140043

Mostly seen as a gimmick. I made the mistake of buying a single core processor (Athlon64 3500+) and regretted it within a year.
When the Core2Quads came out I upgraded straight to one (Q6600) while people were calling quad cores gimmicks and all you needed was dual core.
Zero regrets. It was a great processor.

>> No.71140234


It was a god-sent for the normal desktop world. Multi-tasking became a reality for the masses. You no longer had to worry about making coasters on your CD-R/DVD +/- R burns. You had to invest in HEDT-tier platforms if you wanted SMP support.

>> No.71140252


Wrong, SMP was niche back then even among the enthusiast crowd. You had to get HEDT-tier boards to get SMP support.

>> No.71140302

>How did people react to the original multi core processors?
I liked those PowerPCs, but thought they were too expensive and a hoax.

But once I had a dual CPU G4 [email protected] with tiny caches (coming from a single core 1.4GHz with biggest caches available) I didn't hesitate any longer and went balls deep into the multicore meme and bought an Intel C2D asap. Then upgraded to 4 cores asap one to three years later, I don't recall the exact dates.

Immediately a mediocre reaction ("looks like they ran out of ideas"), then 100% want.

>> No.71140363

>How did people react
Well, the pajeet from anandtech reacted as follows:
> single core is enough for everyone. Buy intel.
Then the other inteltards reacted as follows:
> single core is enough for everyone. Check the pajeet tech dot com
Then everyone else started buying multicore cpus, a.k.a. athlon x2s, and intel reacted as follows:
> dell,hp,lenovo stop using amd's products.
And 2 years later intel managed to glue together 2 p4s and call it pentium D.

>> No.71140364

Nah, any overpriced facebook machine did it. Plug it in and use your one button mouse with many CPUs!

You missed that completely, didn't you? I noticed it's often like that, the windows-only users have actually no idea what's going on the world outside of redmond, whilst Linux/BSD/Mac users have to dual boot for some rare niche sw, which then in return opens up their eyes to things they didn't see yet in their native eco-system.

Somebody should split your head open by force if you have the audacity to call other people wrong when you're wrong yourself. Let's hope some real human bean approaches you soon,

>> No.71140454

Disable AA (looks like GTA 4 and nobody cried about GTA 4) and it'll run on any shitbox. It's one of the most optimized games next to Dota 2. You're full of shit, mate. GTA V even runs on fucking playstations from 15 years ago.


>> No.71140466


>zoomer/latecomer detected

> unironically uses facebook machine as example when SMP hardware and platforms predates failbook by a few decades

> implying that SMP world wasn't enterprise/HEDT-only until the mid-2000s

>> No.71140487

People didn't thought about it much back then.
All they cared for was the massive boost in single threaded performance that this chip did.
And the fact it was so fucking good at caching things, it could fool all the memory speed testing programs.
But i imagine MANY people did enjoyed the fact that the dual core chip was actually able to play audio without getting all the sorts of awful crackling when using integrated audio chip.

>> No.71140508

Get a single core computer then, you will LOVE how windows just go and steal the CPU and your sound breaks, your download gets slow, shit stutter...

>> No.71140526

Phasing out my C2D desktops this year, but will keep the laptops going as far more useful than some think. They can really do any task smoothly if you have a good system for it and yes Windows 7 is one of them if you have 4 GB of ram and SSD.

>> No.71140552

>>>zoomer/latecomer detected
Eh, yes, he said so with the literally first word in this thread. Therefore my wording, which you clearly and unmistakenly understood. No surprise as I didn't stutter. You're confusing me with OP, aren't you? I'm addressing him.

>> implying that SMP world wasn't enterprise/HEDT-only until the mid-2000s
I can see that implication of yours, yes. You even noted it yourself.
To prevent such implications I wrote, even a literal mactoddler could handle and set up multicore systems, already in 2002. That's not mid 2000s, that's early 2000s.
The G3 came out 1999, mate. The G4 came shortly afterwards. That's not mid 2000s, that's early 2000s.

A complete braindead noob could have dual core systems in 2002. You were either too cheap, too dumb or too unaware and now you're pissed as you notice you missed out. Otherwise you wouldn't attack facts.

But you do you. Whatever floats your goat.

>> No.71140813

As with 64 bit architecture the hardware was far ahead of the consumer level software. Took years for programs, especially games, to effectively utilize multiple cores after their introduction. However single core processor development quickly pretty much ceased with the advent of dual cores, and at first people didn't buy dual cores for their dual core capacity but for their increased single core performance over older single core processors.

Also the very first dual core processors were a mixed bag, the AMD Athlon 64 X2 ones in 2005 were eh because the single core AMDs at the time were also eh, and the first Intel Pentium D ones in 2005 were just two Pentium 4 housefires glued together, and things didn't noticably advance until the Core architecture a year later, at which point the first quad cores (albeit expensive) were released only six months later and gradually gained marketshare with increased affordability. So effectively the dual core era of Athlon 64 X2 and Pentium D/Core 2 Duo only even lasted very few years, relatively quickly being replaced by quad cores such as Phenom and Core 2 Quad/Nehalem even in mainstream gaming computers.

tl;dr 2006 to 2009 or so was the only time when dual cores were even relevant, but even then largely for their single core performance, and by the time software caught up to multithreading quad cores were already the big thing.

>> No.71140913


>moving goalposts to used HEDT hardware on fleabay (Dual Slot 1 boards from decommissioned servers/workstation) as "affordable" SMP back in the early 2000s

> implying such systems in that period weren't outclassed by normal desktop SP hardware outside of multi-tasking environments

>> No.71140930

Totally, bro.

>> No.71141236

The conventional wisdom at the time was that quad cores were pointless for a machine intended for gaming because games didn't take advantage of them at all. Then the first Nehalem systems introduced per-core overclocking which mitigated a lot of that and made them around as useful in games as lower core variants so then they started being the go-to option.

>> No.71141480

Intel released the Pentium D to try and foil the Athlon 64 x2's launch. Of course the Pentium D was so bad that it was almost universally panned. It wasn't just that it was a (literal) hot mess it was also a waste of money. At the time there were existing single core Athlon 64s that were better than it and ran cooler and were less expensive. Then of course everyone expected the Athlon 64 x2 would launch soon anyway, which it did, in literally the same month. Pentium D/Extreme was essentially dead on arrival.

>> No.71141511

The two biggest jumps that i'll remember was 64bit and multicores.

Both were pretty lackluster when they were released but have become industry standard.

Take that as you will.

>> No.71141704

They made Windows's shit scheduler bearable if you were multitasking.

Professional software and a handful of games could make use of them; everything else took a while.

>> No.71142142

Subtle Incel shill

>> No.71142194

>t. Adobe developer

>> No.71142297

>Athlon 64 x2 was monolithic.
Not really. Yes, they were rather foresightful about it because they birthed it from multi-CPU enabled Athlon 64s that were previously available on multisocketed motherboards, but in the end the X2 was two Athlon 64s joined on one die. That's also why they're listed with 2x1024kb cache instead of 2048kb. The main reason it worked better than the Pentium D hellfire was that the Athlon 64 was far more energy efficient to begin with, whereas you could barely even cool one Pentium 4, let alone two of them glued together.

>> No.71142361

>Pentium D/Extreme was essentially dead on arrival.
On the upside if you still have a "retro" 775 system around it makes for a neat collector's item these days to rev up every now and then and try to burn your house down with when you're feeling depressed. One of the most interesting sockets of all time probably thanks to its long lifespan, got everything from Pentium 4s to Core 2 Quads, and it was at a time when you could still get motherboards with a plethora of different chipsets.

>> No.71142469

but the housefire meme is on intel now

>> No.71142499
File: 84 KB, 600x428, skulltrail-3q-2.jpg [View same] [iqdb] [saucenao] [google] [report]

>tfw no hoverboat rig

>> No.71142572

>2ghz + 2ghz = 4ghz!!!

>> No.71142609 [DELETED] 
File: 505 KB, 797x443, a10.png [View same] [iqdb] [saucenao] [google] [report]


>> No.71142751

>i think my apache attack helicopter has autism

>> No.71142803
File: 121 KB, 900x735, hovermower.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.71142953

Check this channel out


Dude tests all kinds of high end old chips with newer titles, really interesting to seeing how they hold up

>> No.71142969

I remember my dream rig being one of these with three-way 8800 Ultra SLI

Crysis at 2560x1600, 60FPS!

>> No.71143005

>Core 2 Duo
Good sir, Core 2 Quad was the apex of kickass Intel CPUs back from the 2000s. LGA 775 represent motherfuckers

>> No.71143035


we had multiprocessor mainboards since a lonnnnng time before multicores. It was well received.

>> No.71143086

Pentiums were a meme, but by the Core 2 generation, everyone loved them. Core Duo was wonky.

>> No.71143109

Pentiums were. Until Celerons started to exist.

>> No.71143152


Only for 955X chipsets newer though. 915/925 boards cannot support Core 2 family due to voltage/VRM differences between first generation S775 boards and later units.

>> No.71143217

>"2 cores, each running at 2GHz? I HAVE 4GHZ OF PROCESSING POWER NOOBS"
This. I also thought that for a brief period of time

>> No.71143241

My DFI LanParty NF4 and socket 939 Opteron 175 was the truely the comfiest setup I ever had

>> No.71143302


It was because poor dual core P4 had to share a 800Mhz/1066Mhz FSB on an architecture that thrives on having a fat bus.

Dual core P4 were marginally faster than their SP P4 w/ HT counterparts at multi-tasking.

OTOH K8s had HTT links to play around with. The Athlon X2s were two K8 dies on a single package using a special coherent HTT link to each other.

>> No.71143406

Unironically jelly of you desu.
Those Socket 939 Opterons were great, and the cheap ones often overclocked well.

>> No.71143451

In those days Intel was flexing its monopoly over AMD, so despite the fact that K7 and K8 CPUs were better than Intel on a cost:performance ratio, most people still had shitty Pentium 4s. AMD knew they had to rush out dual core CPUs because this meant that they would be the ones that get to set certain standards for multicore x86 CPUs. It was pretty obvious that there was little in terms of market demand for it at the time, but it wasn't useless. Intel responded with the laughable Pentium D, which was truly the lowest point for Intel. Pentium D didn't do anything and every benchmark labeled it as "well it performs exactly as you'd expect a mobo with two CPUs to perform, until it gets hot, which it does." The chip was basically only useful for encoding, which was already the P4's strong suit.

Not long later, Intel released Core 2 Duo in 2006 and that's when real demand for dual core application support started being demanded from customers. It still took until late 2008/early 2009 for significant support to appear, but the jump from there to quad core support wasn't as delayed.

>> No.71143453
File: 363 KB, 1920x1034, 1558632133383.jpg [View same] [iqdb] [saucenao] [google] [report]

Anon you're so funny. I'll give you all my upboats, and my virginity.

>> No.71143465

ESXi still counts that way because VMware is retarded.

>> No.71143645

I'd get an nForce board anyway for that SLI support.

>> No.71143717

>But once I had a dual CPU G4 [email protected] with tiny caches (coming from a single core 1.4GHz with biggest caches available)
Wat, the G4e had 256K on-die L2 and DP models had shared 2M L3, better than either the 1.4 GHz Athlon or Pentium 4.

>> No.71143765

>You don't need 2 cores, high-speed single core is better
>You don't need 4 cores, high-speed dual-core is better
>You don't need 8 cores, high-speed 4 core is better <- 2017
>You don't need 12-16 cores, high-speed 6-8 cores is better <- we are here

>> No.71143844

Nostalgia. People are afraid to let go of the old and accept the new as they themselves will one day be let go and replaced.

>> No.71143848


I don't think anyone needs 12 cores. People sitting on 8 cores should just wait another year (or not even) for 16.

>> No.71143854

As I remember the big deal was when Intel released the core and core 2 lines and the socket 775. Those were good for everything.

>> No.71143961
File: 409 KB, 1144x888, 7e6.png [View same] [iqdb] [saucenao] [google] [report]

>tfw gentoo user
>tfw core count is going higher
>tfw make -j64

>> No.71144006


Core CPUs were just for laptops. Intel ported them (and Pentium M) to desktop mobos because some customers wanted a CPU that had decent performance but extremely small power draw.

Core 2 was a year later, when Conroe killed AMD.

>> No.71144400

At first it wasn't seen as necessary due to almost all software and gaymen being single threaded. Over time this changed though (within only a couple of years) as things began to become more multi-threaded. I remember buying a Athlon64 Socket 939 single core in 2006 and thinking it would be fine (which it mostly was) but then suddenly went to shit around 2007/8 as things became more multi-threaded.

>> No.71144589

Also don't forget x86-64 is AMDs arch too and it just werks. Intel Itanium x64 was such a mess Intel ended up ditching their own arch for AMDs.

>> No.71144650

>You don't need 2 cores, high-speed single core is better
It's a fact. But we have hit the single core wall. 64 GHz single core CPU with current IPC will be insanely good compare to 64 cores 1 GHz CPU

>> No.71144723

I loved my E8400.
God the only feeling equal to getting it was getting a ssd.

>> No.71144762

They went from coarse-grained multitasking to fine-grained multitasking, that's the main reason single core started to suck

>> No.71144774

Actually I'm not so sure it would be better. It's true that the CPU would in theory be able to do 64 billion operations in a second but what kind of memory would be able to feed that fast enough to make use of all that speed? Not just the memory either there would need to be a super high speed bus between the memory and CPU. And then from the memory to the storage device(s). If not for those things the CPU would stall out whenever it had to access memory and storage.

On top of that increasing the clock speed increases the power usage and temperatures in an almost exponential way. Powering a CPU that runs at 64GHz would require an incredible amount of power unless something fundamentally changes about how CPUs work.

>> No.71144813

fuck off you zoomer faggot
dual cores were a meme for games at first

>> No.71144832

Itanium was alright in high-end systems, but it didn’t totally revolutionized the consumer space like the hype train said it would.

>> No.71144835

when the first multi core cpus came out they were clocked well under single cores crippling them in the current games.

>> No.71144840

>mfw my servers still run on E2160 and E3400
E2160 was the shit. It was 1.8GHz stock, I run it on 3.6GHz on box intel cooler since 2008 and it still just works.
Modern unlocked K-series (CPUs) can suck my nuts.

>> No.71144846

Memory and CPU cannot be divided, if you understanding computer architecture. When I mean 64 GHz CPU, that means it must have 64 GHz internal memory (registers, cache, etc.). Most of our computer are synchronous, not synchronous.

Back in the day when DRAM speed still kept up with CPU speed, we had very little cache, cache is only for latency, not throughput. But eventually, DRAM lagged behind microprocessor, and the cache grew larger.

>> No.71144939

And about powering the CPU, that's mean electrical computer is a dead end, just like mechanical computer before it. We need to think outside of the box and need new way to construct computing element. Quantum computer won't count, because it need quantum comptuting, which is a totally different way of computing than our traditional method.

Speed of light is our greatest enemy. Without breaking it, there is no hope for modern computing.

>> No.71145059

M-Muh Q8200s

>> No.71145084
File: 2 KB, 125x125, 1546365559286s.jpg [View same] [iqdb] [saucenao] [google] [report]

>tfw Gentoo User
>tfw Intel Xeon 7290
>tfw -j288

>> No.71145095

Assuming we have memory fast enough to keep up with the CPU we still wouldn't necessarily have storage drives capable of keeping up. We have that problem right now with hard disks and even solid state drives. On a system with a super fast CPU accesses would still be painful. That's one of the advantages of having multiple cores, those idle times can just stall out on another core while the rest of the system moves on, the end user doesn't necessarily have to notice it. SMT further improves that by allowing the idle core itself to move on to other tasks while keeping the idle thread open in the background.

At the very least I think a CPU with a clock speed that high would at least need some implementation of SMT to deal with tasks that frequently idle.

>> No.71145103

It’s just time to accept the fact that Moore’s law isn’t a law.

>> No.71145198

Moore's law is a real law, if you see the population chart of mankind over history. There would be thousands of years of almost no major populations increase and some points, the populations literally exploded. 19th century and 20th century was when productivity exploded exponentially, and now is the plateau of normalcy. Until a new technological revolution, we expect no major advancement in computing. Moore's law is not wrong, but we're not ready for a new period of Moore's law yet, it could take decades, centuries, or even millennia, no one could know yet.

>> No.71145280

You're just thinking 64 GHz in term of our current technology, of course it's fucking impossible. 64 GHz means the clock cycle is 0.015625 ns. In that kind of time, the signal need to travel all the chip, which means the longest possible path line is 4.6875 mm, so how could our 7nm tech achieve it?

>> No.71145447

Same as today. Which tells you a lot.
Gamers, innane and dumb as always failed to see into the future and deemed it as a gimmick (or rather, inferior). They were soon dragged kicking and screaming into the current year.
Enthusiast, professionals, whatever you want to call us saw it as great - like a dual CPU workstation cheaper and smaller.

>And about powering the CPU, that's mean electrical computer is a dead end
No. Single core is dead, that's all. They could only lazily increase frequency in the past because early CPUs were so weak that they hadn't hit the power wall yet. Multicore is the solution to the power wall, and it has been solving it since 2005.

Then we ran into the "die size wall" or "yeilds wall", which AMDs chiplet design efficiently solves. People are successfully thinking outside the box - they just aren't thinking inside your single core box.

>> No.71145749

I'd still be using my Q6600 if the mobo didn't shit the bed. That thing overclocked nicely.

>> No.71145951

lmao what

you mean all the interesting/useful shit IS multithreaded

brainlet corelet nocodelet shill detected

really has to be a troll post.

>> No.71146155


This isn't true. Gamers all had dual core CPUs a couple of years before devs supported them. Conroe could not be produced fast enough to meet sales.

>> No.71146496

Pentium D was fucking complete shit

>> No.71146951
File: 1.47 MB, 3472x2604, 100_3407.jpg [View same] [iqdb] [saucenao] [google] [report]


Um, yeah, we did have SMP machines.

I've had at least one dual proc machine in my house since my old IBM PC 365. Dual Pentium Pro 200's, eventually upgraded with 333MHz Pentium II Overdrive processors. Abit had one of the best "enthusuiast" dual proc boards during the P3 era (the BP6, which I owned). There were plenty of dual Slot1 and Socket370 boards in those days. My last dual P3 was a dual 1.4GHz Tulatin machine.

Pic related is my one remaining dual proc machine left. It's a dual 1.2GHz Athlon MP with
a Tyan Tiger MP motherboard, w/2GB RAM, 64MB Geforce2MX, random CMedia audio card, and a Netgear gig ethernet card. It's my old NetBSD machine. I've disassembled it and the board and procs are sitting in a drawer next to me right now due to leaking caps on the board. I'll get around to fixing it and I'll probably upgrade the procs to Palomino MP's and maybe upgrade the video card to something that can play 1080p video.

>> No.71146972
File: 76 KB, 674x532, Quadro_fx1800_tesla.png [View same] [iqdb] [saucenao] [google] [report]

> intel managed to glue together 2 p4s and call it pentium D.

>> No.71147004

>Pentium D was fucking complete shit
It's almost like I responded to a post that implied that saying the intel dualcore CPUs were shit at first is a lie or something.

>> No.71147115

>first 64 bit
I think I didn't forget

>> No.71147352
File: 111 KB, 810x615, x3220_c2q6600.png [View same] [iqdb] [saucenao] [google] [report]


I'm still using mine. I'm posting from it and it's my daily driver for the most part (I have a haswell i5 that I use for my client work but that's it).

>> No.71147499

now they have [insert]-Lake "90W" housefires

>> No.71147540

>All the actually interesting shit cannot be parallelized well.
I guess just games are interesting shit to you.

>> No.71147560

It was like discovering B:, if you know what I mean.

>> No.71147614

Some of us did, but he's quite right to say it was niche in a time when mainstream consumer operating systems weren't even SMP-aware, few home users, enthusiast or otherwise were blowing cash on multi-socket machines.

>> No.71147644

Nearly every adult I knew who had rigs they built were running dual cpus

>> No.71147692
File: 32 KB, 500x387, 1390265771875.jpg [View same] [iqdb] [saucenao] [google] [report]

>How did people react to the original multi core processors?
As an enthusiast, it was a nice to have. It was certainly cheaper than dual CPU boards, which were really pro grade and super expensive.
>Were they immediately well received or seen as more of a gimmick?
Well received as it was a nice boost to multi-tasking even if the OS wasn't fully optimized. Most software at the time was still single thread oriented with devs that tuned for it. Most OSes at the time could schedule for multiple CPUs but weren't really great at it.

>> No.71147844


IIRC Toms Hardware ran an article in the late 90's where they overclocked a pair of celerons to 500MHz each and declared "1ghz processing powar!!!" or something to that effect. I'd have to do mah googles to find the article, but it was a pretty big deal back then.

>> No.71147923

And everyone I knew was using a second-hand Sun or SGI. Funny how anecdotes are.

>> No.71147965

cow2beef-posting mostly.

>> No.71148115


dude, when it was found that Quake 3 Arena was SMP capable every guy I knew running a rig capable of playing it went out and copped a BP6 or similar board and a pair of Celeron 333's or 366's to run 'em at 500MHz/550MHz and nice AGP video card.

I had it 'cause I always had an NT4 machine alongside my Win95/98 box. When I finally got into Linux and FreeBSD it made even more sense to keep an SMP machine around.

>> No.71148131

*raises hand*

I had an Indy and an SS20 along with my dual proc machines.


>> No.71148229


You don't matter, he said SMP was niche and it was. You'd have to go to a lan with hundreds of machines to find more than a handful of dual CPU machines, and that was the enthusiast crowd.

>> No.71148562

is that fucking world of warcraft mists of pandaria?

>> No.71148680


At the time most OS'es were quite optimized for multiple processors. Windows had been SMP capable since NT, and with Win2000 support was better but you were limited to 2 procs/cores under Win2000 Workstation. With XP you could get quad core or greater (my workstation at the office from '07 though 2011 was a dual quad core Dell Precision 690 running XP). There was a test run between XP, Vista, and Win7 about 10 years ago and 2-4 cores, XP was the clear, hands down winner. However, it was stated that as things progress, and they optimized more, and the number of cores increased, the two newer OS'es (at the time) would be better at at it.

But... yeah.. XP spanked them both up.

found the article too: https://www.infoworld.com/article/2674526/the-generation-gap--windows-on-multicore.html

>> No.71148888


It should be noted that that was a release candidate of Win7. Win7's performance by 2010 had significantly improved due to driver optimizations and general updates.

>> No.71149857


It a small niche back then. I never stated it was non-existent. Dual Socket A and Dual Slot 1 boards were popular choices for small number of enthusaist who went the extra mile for SMP. The aforementioned boards were would currently be regarded as HEDT-tier boards.

It was niche as current Threadripper/Skylake-X family stuff.

SMP was practically tiny outside of enterprise and SMB world until the era of multi-core chips.

>> No.71149964

>for games

>> No.71150028

My first dual core was Yonah, T2300/2400.

It was clocked fairly low (1.66, 1.83) but since it was Pentium M architecture it was the among the fastest CPUs at the time, even being a mobile chip. I had that Tohisba U200 through the end of 2009 when I replaced it.

>> No.71150051

>You don't need more than 1 core!

>> No.71150265

Leaving aside advanced issues such as cache coherency, each core can only use half of the 800 MT/s FSB bandwidth when under heavy load.


>> No.71150308
File: 42 KB, 285x287, 1345217491389.jpg [View same] [iqdb] [saucenao] [google] [report]

q9550 reporting in. Looking forward to upgrading to 3900x later this year, I'd imagine it will feel like switching from a tractor to a Ferrari.

>> No.71150383
File: 894 KB, 1384x1880, 3-4.jpg [View same] [iqdb] [saucenao] [google] [report]

Nice nazi quads, checked!

>> No.71150384

They were still slower than my Pentium 4 3.4ghz processor.

>> No.71150405

This allowed for multi-track shitposting.

>> No.71150413

Shit lake 90++++++ W TDP

>> No.71150545

Actual non-trivial code has irremovable control and data dependencies. It's what separates it from dumb parallelized biology.

>> No.71150806

It was split. Anyone with actual knowledge of computing saw the value in it. Most people saw it as a gimmick. Of course this wasn't unreasonable since very little consumer software at the time was capable of running on multiple CPU cores.

>> No.71152210

Okay zoomer, let me school you on something.
No-one saw them as a gimmick because whenever any application hit 100% or there was a memory leak your whole PC would crawl to an actual stand still. Sure it took a while for applications to make use of the other cores but in general it was great to not be a corelet.

Dual core processors and HT were seen as a god send. Anyone remembering them as a meme is retarded.

Name (leave empty)
Comment (leave empty)
Password [?]Password used for file deletion.