What is Digg?Sponsored by Travelzoo
Take Advantage of Ridiculously Low Holiday Airfares view!
travelzoo.com - Flights $52 and up for Thanksgiving, Christmas & New Year. But move on it now.
58 Comments
- retral, on 10/12/2007, -3/+18Hardware 'dying' is overrated; heck, even socket A chips/motherboards are still being rather widely sold/used, so, at least in my opinion, they're not dead yet, but supposedly they 'died' several years ago.. And how is socket 754 old and low end? It might not be brand spanking new, and it might not be cutting edge, but it's used quite widely for mobile turion 64 laptops which tend to be a decent balance between power & battery life (not to mention 64-bit). That's really the only way it might 'outlast' the other two if you think about it..
Dugg nevertheless. - 1ivewire, on 10/12/2007, -0/+10In that case scott, I have some old "hardware" you might be interested in.
- inactive, on 10/12/2007, -0/+10Don't Freak Out People. Hardware can never truly die, only lose support.
- inactive, on 10/12/2007, -0/+9Theres a LOT more than just processor speed at play when you are talking about the speed of an entire system.
You could have the fastest Opteron with 256Mb RAM with bad timings and it would have its pants beaten off by a lowly Athlon2400 with heaps of dual channel ram configured correctly and a good video card sitting in a good quality motherboard. - AXNJAXN, on 10/12/2007, -1/+8I'm still hesitating to believe ANYTHING about it until April 2nd...
- pwncore, on 10/12/2007, -0/+6I still run a socket 754 with an AMD Athlon 64 3400+. I built it two years ago and the only thing I've upgraded was the video card. I first built it with a GeForce 5700 and later replaced it with a 6600 GT. It still runs games at the highest settings without a hitch. I decided to wait out the socket 939 craze because I saw no point in spending $400+ on a new mobo, CPU, and PCI-E video card to play the same games I'm already playing without any problems. People need to stop buying what the coolest and newest is and instead consider how much they really need it. My next system will be a AM2 socket system and that $400 I would have used to upgrade before has grown, allowing me to build an even better AM2 system.
- Morca007, on 10/12/2007, -0/+5Whatever, now all this means is when I want to upgrade my s939 it'll be cheaper. *rolleyes
- matts0344, on 10/12/2007, -1/+5I've had some die on me before.
- mirunit, on 10/12/2007, -1/+5Well Socket A is still here and still works fine, Socket 754 and 939 are pretty much the same with exception to duel channel memory and power consumption. To the poster who said all amd PCs above 500 are 939, that is false. Socket 754 is still being sold because it is an efficient alternative to 939 and still supports AGP because a large portion of the market now does not have PCIE.
- brownb2, on 10/12/2007, -2/+6Those "inefficent" programming languages allow for higher security, lower likelihood of programming errors, and better program isolation meaning products are faster to market due to less testing required for memory leaks, buffer overflows etc. Given the complexity and size of modern software I would consider the benefits of C# and Java to be more than worthy of a few extra clock cycles. At worst we're talking a few extra seconds on startup, and an extra 16-64 megs of RAM for the VM (for a desktop, embedded systems have much much smaller overheads). Now thats hardly breaking the bank, and it certainly not driving processor sales or warranting your suggestion of bloat. Or do you complain because processors are now dual core, will soon have CPU level virtualisation and are better than older models, which required minimum ~400 mhz processor to decode mpegs streams whilst multitasking, or to re-encode your DVDs for PDA would take a day, couldn't play the latest uber realistic flight sim and certainly couldn't do simple speech to text?
A long story short - the latest technology always brings new benefits, the quest for speed is driven by the desire to produce software current machines are incapable or barely capable of. That and the I've got a bigger willy syndrome of young men. - Titan615, on 10/12/2007, -0/+3This isn't news to me, they have been discussing the AM2 on Anandtech for a while now and I can't wait for the upgrade. I mean, all they are doing with the AM2 is they are going to start selling 65nm core versions of the AMD 64 line and the X2 line. Also this socket will support the X2 5000 and the FX-62. The 939 is not going away, you can still get really high quality processors for it like the FX-60 etc, its just that the newest chips aren't going to be on the socket 939.
- veritech, on 10/12/2007, -3/+6I got my overclocked skt A 2600+ right here, it works, why change it? anyone who's into hardware knew that 939 was dying since summer last year, and that 754 pretty much dead with the widespread use of 939.
Anyways a lot of the software we use doesn't take full advantage of older systems never mind newer ones, that added to the bloat of a certain "modern" os + plus some of the more inefficent high level programming languages, and you have our current quest for speed. - pwncore, on 10/12/2007, -0/+3If AMD stuck with the 939 you'd be complaining 6 months from now that they don't have DDR 2 support yet and need to get with the times. Technology evolves, get over it.
- inactive, on 10/12/2007, -2/+4@veritech
If you look almost any source code for an application, the amount of inefficiency caused by the programmer will outweigh losses from the language by leaps and bounds. Many high level programming languages (e.g. Java, C++, Python) are theoretically only 5%-10% less efficient than programming in Assembly. On the practical side, programming at a lower level has an average of 400% increase in errors/memory leaks and frequently has worse performance than a higher level compiled language because a higher level language has more metadata available, and can put the programmer's method invocations in context. A lower-level language has no context for what the programmer is trying to accomplish, and so can only follow his instructions one step at a time until completed.
You can think of it this way, if I give you step by step instructions: get your keys, go to the front door, open your front door, exit the front door, close the front door, go down the stairs, open the car door, get in the car... etc. All you can do is just follow the steps as they are given, you can very well predict where all this is going, there's no context to each step.
If, however, you're told "Go visit your aunt in Kentucky", you can figure out the intermediate steps for yourself, and definitely do them with greater efficiency than someone giving you step by step instructions, discarding redundant steps and combining multiple steps into single actions, because you have more metadata to work with and you know the goal. - xpantz, on 10/12/2007, -0/+2M2 DDR2... its all very nice. But don't upgrade until you can see a serious performance increase.
- rekka, on 10/12/2007, -0/+2@brownb2
Good point, but your last sentence was the most indisputable. - Splitt3rxx, on 10/12/2007, -0/+2this is *****, I just bought an hp with socket 939 because I thought it would last longer than 754 or any of intels crap.
- jamelt, on 10/12/2007, -0/+2WHO CARES!, I already have my processor, if you get a processor after AM2 is out great for you big deal?
- thoffmeyer, on 10/12/2007, -1/+3I hope the x2's for the am2 are cheap, because I was going to buy one for 939, and now there going to be dead :(
- joesnow, on 10/12/2007, -0/+2some of you seem like you believe that you'll be unable to upgrade ever again w/ that socket, you have a little over a year to decide to upgrade again, that's awhile in regard to CPU tech, and that doesn't mean that the same chips won't be out 3 years from now, they just won't be manufactured brand new anymore..... hell I know some stores that still have brand new PIII chips in sealed boxes..celerons, thunderbirds, etc.. you can still get'm very easily. Beyond that, this is nothing new, this is how just about every technology has gone in and out, only difference is, this particular story got digged so it makes you think about it, as opposed to just going with the flow and not worrying.
I got a 939, expecting this socket to only last me one more upgrade of CPU, beyond that I expect to get a whole new motherboard/CPU-type/memory-type when I upgrade.....as it's always been. - Brennan, on 10/12/2007, -0/+2Haha score, I just sold my 939 desktop for a laptop with a 754 because I needed the portability and it was cheap. I hope they release some cool 754 processors.
- aristofeles, on 10/12/2007, -0/+2Doesn't look like, but I still hope...
- SoulMaster2, on 10/12/2007, -1/+3@adml_shake
Are both CPUs using the same core? - sw96, on 10/12/2007, -0/+2Socket 754 = no dual channel, mostly AGP mobos.
- veritech, on 10/12/2007, -0/+2754 is even bigger than 939 for one reason, IGP, OEM's love this stuff like, babies love milk. Drives me nuts, but everyones gotta earn
- weiran, on 10/12/2007, -0/+2I'm still considering whether to get a Socket 939 Opteron 165 now or wait for AM2, even though it would cost a lot more to get me the same spec (as I can't use my current DDR1 RAM).
- khyberkitsune, on 10/12/2007, -0/+1"Currently, socket 939 is home to AMD's Athlon 64, Athlon 64 X2 and Athlon 64 FX processors while socket 754 is for Sempron."
*****. I'm running a 754 Athlon 64 3000+ Too bad I can't bury as inaccurate twice. - marmaduke, on 10/12/2007, -2/+3I just think that its very annoying that they just started making 939's a short while ago. I got my current setup so that I could expand. Just wait, soon PCI-e cards will go out.
- brownb2, on 10/12/2007, -0/+1@StringCheesian
I think you'll find the 5-10% quote isn't mine - its raisenero's so I can't comment.
Regarding your other point - my thoughts as a developer of 8 years is that lower level more efficient languages have their place with older systems, or limited systems such as embedded, PDAs and so forth. I regularly work on projects with thousands and thousands of lines of code, I QA as much of my team's work as possible but I cannot catch all bugs. Its not just the programmer's skill - everyone makes mistakes, and any system that is *guaranteed* to remove certain types of bugs is a welcome addition (i.e. a JVM sandbox), unlike QA which may miss them. You may disagree with the time to market example I raised, but investors and shareholders would not, likewise on today's high performance machines the difference is negligable. So I agree that lower level languages C, assembler have uses, just less so nowadays, although one main reason I believe Java/C# languages are not main stream published (i.e. games/apps) at the moment is simply one of source code piracy, C# Java etc can all be reverse engineered whereas C (for example) is extremely difficult. I cite Microsoft's Vista OS as an example of good intentions but either lack of good C# developers or threat of piracy preventing the transition. As an aside Java is fast enough to allow emulation in a browser and openGL 3D games so speed on games is not an issue. I hope you see the angle I'm coming from. - Abyss_908, on 10/12/2007, -0/+1That sucks. Although it is understandable I guess. 754 must be must cheaper to make and by the time dual core becomes mainstream there will be another socket and AM2 can be the budget one. Sucks for me though I just build a 939 system.
- khyberkitsune, on 10/12/2007, -0/+1Buried as Inaccurate.
AM2 is already in use in HP Laptops. I see them everyday at work, they're already on the market. - diggnationdevon, on 10/12/2007, -0/+1This makes perfect sense because the 754 is more budget-friendly and the new M2 socket is the successor to the 939. But this is great, the new socket will be fantastic.
- redsrule2500, on 10/12/2007, -0/+1By 'die' they mean not be used in brand new computers?
- schaver, on 10/12/2007, -0/+1About time! This hardware has been out for. . . MORE THAN TWO WEEKS?! And AMD's just now getting rid of it?? Come on guys, pick up the pace! Socket 939 currently supports the X2 processors that AMD is releasing and that's good enough for me for right now.
@retral
Good comment on your part - pillfred, on 10/12/2007, -0/+1***** I'm still gonna upgrade my 754 here in a few months to pretty much the specs Pwncore is running for the same reasons. Chip and vid card $289. It's all hype. Just because the FX-61 is coming out doesn't mean that my buddies FX-53 is suddenly a ***** chip.
- khyberkitsune, on 10/12/2007, -0/+1The AM2 socket will allow for future upgrades to take better advantage of the HyperTransport bus.
- ACalcutt, on 10/12/2007, -0/+1i really find that hard to beleive seeing 939 is the better technology
- JamesGlover, on 10/12/2007, -0/+1Shame, looks like I will be doing a Mobo replacement when I come to upgrade my processor. I built a 939 system back in early August hoping that it would last a while and the cost difference was minimal. (Same reason I went PCI-E and SATA) Having said that I may be able to squirm in with a top-end end of the line, but we'll see.
- kodek, on 10/12/2007, -0/+1Well, what other options did you have? :-)
- inactive, on 10/12/2007, -0/+1go ddr2 thats what im going to do!
- dylanrjones, on 10/12/2007, -0/+1Does anyone know anything about price differences between the 939 hardware (mobo and cpu) vs the new ones? I was just about to upgrade and I picked out all the pieces I want, but am now worried about it not being as upgradeable. I also don't want to spend a mint. If the AM2 stuff is about the same price, I'll wait. If it's going to be expensive, screw it, I'll get a few years out of the 939.
- beni, on 10/12/2007, -1/+1I thought they were keeping 939 longer than 754.
- RobotCitizen, on 10/12/2007, -2/+2Planned obsolescence blows. Such a transparent scam.
- StringCheesian, on 10/12/2007, -1/+1@brownb2
You say decreased time to market and reduced cost are worth a few CPU cycles and increased mem usage - I take the opposite view: I think the extra cost of hiring people who know what they're doing and the extra time taken by programming carefully is well worth it for lightning fast tiny-executable small memory footprint apps that look as good as and do everything their big slow counterparts can do, and with friendlier system requirements to boot!
Your priorities are not the only valid priorities. I would pay more for fast efficient software, even if it had to be late and more expensive to be as stable.
Also, you think Python is theoretically only 5-10% less efficient than C or assembler? What theory is that? In practice, I've never seen it benchmark that well. And I seriously doubt Java or any other VM/bytecode/interpretted language could ever match the speed of carefully coded assembler with good use of SIMD instructions. - rekka, on 10/12/2007, -1/+1sader; You're 12, right?
- sader1264, on 10/12/2007, -2/+1Actually I'm only 4......so bite me.
- Pestilence, on 10/12/2007, -4/+3April Fools?
- fredinator, on 10/12/2007, -2/+1I guess i'm gunna have to wait a few months upgrade my computer then if i want an amd processor
- fuelvolts, on 10/12/2007, -2/+1ahem.....Econ 101:
less supply = more $$$
take a look at socket A athlon xp - they are more expensive than a faster s754 sempron or a s939 venice 3000
for a 3 yr old chip! -
Show 51 - 59 of 59 discussions
© Digg Inc. 2009
— Content created and posted by Digg users is