AMD's Phenom II 965, 3.4GHz, 140 Watts, $245

Vigile writes "While AMD does not have the muscle to push around the i7, they certainly have the ability to give the older and more common Core 2 Quads a run for their money. With the release of the Phenom II X4 965, AMD further attempts to dethrone the Core 2 Quad as the premier midrange CPU offering. While it may not be a world-beater by any stretch of the imagination, it certainly is catching Intel's attention in the breadbasket of the CPU market. The X4 965 is the fastest clocked processor that AMD has ever produced, much less shipped in mass quantities. While the speed bump is appreciated, the cost in terms of power and heat will make the introduction of the X4 965 problematic for some. Many of us thought that we would never see another 140 watt processor (as the Phenom 9950 was), but unfortunately those days are back. Still, AMD offers a compelling part at a reasonable price, and their motherboard support for this new 140 watt processor is robust."

Comments

as the power station would say...

[BobZee1]

some like it hot

FAIL

[Anonymous Coward]

I'm running a Q9550 at 3.4 Ghz right now (with the ability to go much higher) and mine only uses 95 watts.

AMD has a long ways to go to get back in the game. I can't imagine craptacular ideas like purchasing ATI are helping.

Re:FAIL

[avandesande]

219 dollars at newegg

Re:FAIL

[Anonymous Coward]

I paid $260 (with discounts) for my Q9550 nearly a year ago. It's only $220 retail now, without any discounts.

Also consider Intel has a 65W version of the Q9550.

I can't see how this AMD is any good.

Re:FAIL

[Anonymous Coward]

Intel and AMD release different numbers for their CPU's power consumption. Intel gives an average and AMD gives a maximum. They're not comparible. In real world testing, the X4 965 uses slightly less power at idle and slightly more power at full load than a stock Q9550.

Re:FAIL

[Anonymous Coward]

It has been fairly well proven that AMD is full of it in that regard. Just search and you will find lots of tests comparing the actual power usage and Intel always comes out on top.

Re:FAIL

[A Friendly Troll]

Intel and AMD release different numbers for their CPU's power consumption. Intel gives an average and AMD gives a maximum. They're not comparible. In real world testing, the X4 965 uses slightly less power at idle and slightly more power at full load than a stock Q9550.

Sadly, that is incorrect.

http://www.xbitlabs.com/articles/cpu/display/phenom-ii-x4-965_4.html#sect0

While idle is comparable (Intel has a slight lead), full load most definitely isn't - 75% extra power consumption (which amounts to over 60W!) for the X4 965 over a Q9550 is far from "slightly more".

Re:FAIL

[lalena]

In many of these tests, they measure the power used to perform a certain task. Even if the Intel consumes 10% more power, if it gets done 25% faster then it consumes less power overall.

Re:FAIL

[avandesande]

It certainly appears that the Q9550 is in a sweet spot for low cost (including platform), power draw and performance.

Re:FAIL

[Anonymous Coward]

Not a chance. If you're running it overclocked you're way over that 95 watts.

Re:FAIL

[mdm-adph]

You were running 95 watts at stock (2.83 GHz). You're way, way over that by now!

Re:FAIL

[ThePhilips]

And motherboard?

At least before, motherboards for AMD CPUs were cheaper, often nullifying short-lived Intel's CPU price advantage.

Though power consumption alone turns me off.

Re:FAIL

[Moryath]

Speccing systems over the years, I'm amazed that people don't go more in-depth.

Every time I spec, I find that I can get an equivalently-performing AMD chip (plus motherboard and RAM and a nice, gigantic Zalman 9700-style heatsink) for ~$50 less than an equivalent Intel rig.

The Intel rig uses a stock HSF (meaning higher running temps and more risk) and still tends to cost ~$40 more each for the motherboard and processor. I've tried to see what I could do to get it to equal out, but there just aren't Intel boards from companies with anything approaching a decent reputation for cheaper.

Yes, Intel's been pushing the high-end scale in recent years. If you want to blow upwards of $400 on a processor, go ahead. In my price/performance range, I get more out of my money going AMD.

Re:FAIL

[Anonymous Coward]

Tech Report tells a different story:

http://techreport.com/articles.x/17402/11

A story that says that you can't necessarily compare like-for-like when using different motherboards. At any rate, xbit seems to have grabbed a pretty power hungry MB for their testing. Looking back at the TR article, the Q9550 numbers are probably a little high, though, as Intel has told them to underclock a different processor for their Q9550 numbers.

In either case, you'd normally want a MB/Processor combo that was lower at Idle than at full load anyway, as most people will not be pegging their CPU all the time.

Re:FAIL

[ThePhilips]

Let's wait a month - and see price then. AMD drops prices for their CPUs quite soon.

Still, 140W is bit over the top. Considering that workstation market is dominated by Intel CPUs which have now lead in both performance and power consumption, the CPU have few niches left.

Re:FAIL

[Pentium100]

The difference is that the Phenom 2 965 is designed to run at this frequency, while you are overclocking your CPU. It's great when overclocking works, but not all CPUs may be able to do it (I cannot go to a store and buy this CPU thinking that I would run it at 3.4GHz - I may get a CPU that runs OK at the specified frequency but cannot be overclocked much).

And the power consumption isn't that big, my dual Opteron 270 PC probably uses more power (each CPU has TDP of 95W).

Re:FAIL

[Anonymous Coward]

Uh, worse case would be a linear increase in power usage. In that case you're looking at a 20% increase in clock which equates to 115 watts. Still way less than the AMD and I doubt the power usage increases linearly anyway so the Intel is probably even better still.

Re:FAIL

[eric_brissette]

The last power consumption comparison I saw showed that the AMD processor consumed 26% more power on average, and a full 50% more power to complete the same benchmark. Either way you slice it, AMD processors seem to use more power.. but I would say that the 50% figure gives you a better picture of real-world performance.

Re:FAIL

[randyest]

How much usage is needed before your extra electricity bill costs eat up that $50 savings? 60W extra at 8 hours a day 365 days per year and $0.10 per kWh is $17.50 for just one year.

Re:FAIL

[Moryath]

The power consumption profile is no different between the two machines I spec'ed two weeks ago.

And you can save far more on your electric bill by keeping your A/C on a reasonable temperature, doing heavy-duty things on off-cycle hours with a smart meter, not leaving lights on everywhere, switching to CFL bulbs (LED's are still a tad expensive), and making sure that your windows/doors are properly sealed.

Oh, and let's not forget: invest in some proper rechargeable batteries for your Wiimote.

Re:FAIL

[Nuno Sa]

Sadly, that is incorrect.

http://techreport.com/articles.x/17402/11

"Interestingly enough, the systems based on Phenom II quad-cores (including the X4 965) draw quite a bit less power at idle than our Q9550-based test system."

"That said, the X4 965-based system draws only 15W more than the Q9550-based one. The gap between the Q9550- and X4 965-based systems is thus smaller than the processors' TDP ratings alone suggest. [In full load]"

"By virtue of its lower system power draw at idle and its ability to finish the rendering task sooner, the Phenom II X4 965 fares better than the Q9550 in our two most important measures of energy efficiency."

Re:FAIL

[Rudeboy777]

The difference is not quite so stark at Anandtech, I wonder what the difference is between their tests?

http://anandtech.com/cpuchipsets/showdoc.aspx?i=3619&p=8

Re:FAIL

[A Friendly Troll]

Tech Report tells a different story: ... A story that says that you can't necessarily compare like-for-like when using different motherboards.

Which is why X-bit Labs did this:

To get a better idea of the situation, we performed a separate Phenom II X4 965 power consumption test under heavy load when none of the other system components are taken into account. To be more exact, we measured the consumption along the 12 V power line connected directly to the processor voltage regulator on the mainboard. In other words, this measurement method didnt take into account the efficiency of the voltage regulator circuitry. ... and got 84W for the Q9550 and 147.6W for the X4 965. Granted, like they said in the last sentence, this doesn't take VRM efficiency into account, and it might as well be that those on the AMD motherboard were woefully inefficient and the CPU itself uses plenty less watts than measured. Still, this is 75% extra, and the VRMs cannot be the main reason for it.

They do similar measurements for graphics cards, too, because measuring total system power consumption from the wall can only get you so far.

I'd personally be more inclined to trust the Ukrainians.

Re:FAIL

[sexconker]

Worst case is not a linear increase in power usage, dipshit.

You've got to crank up your voltages to get that high as well.
Then your fans run more.

It adds up fast.

Re:FAIL

[A Friendly Troll]

Not necessarily. My CPU is overclocked by 33% and undervolted by 15%. It can do a 50% overclock with stock voltage.

Re:FAIL

[mcvos]

That's the best case scenario, actually. Overclocking may require increasing the voltage, which means your power consumption goes way up.

Re:FAIL

[evanbd]

Generally speaking, CMOS power consumption is the result of charging and discharging gate capacitors. The charge required to fully charge the gate grows with the voltage; charge times frequency is current. Voltage times current is power. So, as you raise the voltage, the current consumption grows linearly, and the power consumption quadratically, at a fixed frequency. Once you reach the frequency limit of the chip without raising the voltage, further frequency increases are normally proportional to voltage. In other words, once you have to start raising the voltage, power consumption tends to rise with the cube of frequency.

Re:FAIL

[hattig]

Hmm, I'd trust the Tech Report figures more, which support the post that you replied to.

Of course individual chips differ, platforms tested upon differ, and in the end a few percent either way is neither here nor there.

Re:FAIL

[SavTM]

I think it really depends on what the application is. Most people have idle cores the majority of the time. I tend to aim for lower clockspeeds and wattage, myself. You're right that it's an advantage to Intel over the long term, excepting that there will be 32nm socket AM3 processors in 24 months or less.

Of course, schedules do slip and blunders do occur as well. Hopefully AMD manages to handle all its debts to continue operations, because the market really does suffer without competition.

Re:FAIL

[Aphonia]

they've been reasonably close for the core 2 space and the AM2 space for quite a while.

Re:FAIL

[SBrach]

Oh, and let's not forget: invest in some proper rechargeable batteries for your Wiimote.

How would that decrease my electric bill compared to using normal AA's.

Re:FAIL

[Moryath]

Actually, it puts a very small (VERY small) increase. But you don't waste the money on AA's. You'll save ~$60/year at least.

Re:FAIL

[Anonymous Coward]

Which is why X-bit Labs did this:

To get a better idea of the situation, we performed a separate Phenom II X4 965 power consumption test under heavy load when none of the other system components are taken into account. To be more exact, we measured the consumption along the 12 V power line connected directly to the processor voltage regulator on the mainboard. In other words, this measurement method didnt take into account the efficiency of the voltage regulator circuitry. ... and got 84W for the Q9550 and 147.6W for the X4 965. Granted, like they said in the last sentence, this doesn't take VRM efficiency into account, and it might as well be that those on the AMD motherboard were woefully inefficient and the CPU itself uses plenty less watts than measured. Still, this is 75% extra, and the VRMs cannot be the main reason for it.

It doesn't take a lot of thinking to see that the Q9550 power draw is bunk. Their fully loaded system power for the machine was 208 Watts. Their Idle load was 120 Watts. Either the Q9550 processor supplies 4 extra Watts to the motherboard at idle or something fishy is going on there.

There's also something horribly wrong with the 147 Watt measurement for the AMD processor. That's 7 Watts over it's absolute maximum design rating which would put it over and above what the Motherboard manufacturer's are supposed to design for. Add in that people have been successfully overclocking this same chip by quite a bit and it doesn't take much thought to see that they're way over what the chip is actually pulling.

In short, X-Bit has power numbers that are well under the real Q9550 numbers and well over the real X4 965 numbers which would make them seem to be fairly untrustworthy in my opinion.

Re:FAIL

[Anonymous Coward]

Heh, went back after my previous post and looked again:

AMD Idle: 126
AMD load: 279

Maximum load the processor should put out: 140 Watts
(this is the design reference for MB manufacturers)
Difference between x-bit numbers: 153 Watts

Which, if they were truly measuring CPU load only would mean that the CPU would be pulling much more than 153 Watts at full load as the processor would still be using power at idle. They're basically claiming the processor is using somewhere around 170-180 Watts.

X-Bit's methodology is bunk and I would not be inclined to trust them.

Re:FAIL

[warrior]

For a while it will be proportional to (V^2)*f. However, transistor leakage is exponential with respect to Vds due to DIBL (drain-induced barrier lowering) and the resulting decrease in threshold voltage. Raising the voltage too much is a double-whammy WRT power consumption.

Re:FAIL

[SBrach]

People pay for batteries???!!!!???!111!!!1??11!!!

Re:FAIL

[obarthelemy]

to me, AMD has a price advantage if you take motherboards into account, in the low and mid range.

the CPUs seem to be priced about the same, but the boards are cheaper. I don't know why.

Re:FAIL

[Anonymous Coward]

I'm running a Q9550 at 3.4 Ghz right now (with the ability to go much higher) and mine only uses 95 watts.

AMD has a long ways to go to get back in the game. I can't imagine craptacular ideas like purchasing ATI are helping.

Yours was 95stock... 3.4 Ghz as 95w? I call bullshit bullshit bullshit bullshit bullshit bullshit bullshit

Re:FAIL

[A Friendly Troll]

It doesn't take a lot of thinking to see that the Q9550 power draw is bunk. Their fully loaded system power for the machine was 208 Watts. Their Idle load was 120 Watts. Either the Q9550 processor supplies 4 extra Watts to the motherboard at idle or something fishy is going on there.

Nothing fishy going on. Power supplies and VRMs on motherboard have efficiencies below 100%.

If the PSU has 70% efficiency at 70W load, which isn't unheard of, it will draw 70/0.7 = 100W from the wall. If it has 80% efficiency at 180W load, it will draw 225W from the wall. The difference measured from the wall in the two scenarios is thus 225-100=125W, while the component load increased by 180-70=110W in this example, for a "mysterious" difference of 15W.

Also, don't forget that you can never only stress the CPU; the motherboard takes its power, too. Higher temperature also increases power consumption.

Nothing is linear.

There's also something horribly wrong with the 147 Watt measurement for the AMD processor. That's 7 Watts over it's absolute maximum design rating which would put it over and above what the Motherboard manufacturer's are supposed to design for.

Motherboard manufacturers *always* design for higher loads because of overclockers and future CPUs. It is also quite possible that other AMD motherboards have more efficient VRMs, so an X4 965 using 147W on that motherboard could use 120W on another.

In short, X-Bit has power numbers that are well under the real Q9550 numbers and well over the real X4 965 numbers which would make them seem to be fairly untrustworthy in my opinion.

Not at all. They are the only ones to have tested the true CPU power consumption - granted, before the VRMs kick in, as there's no other way of doing it. Other sites only measured total system consumption from the wall, which is far more inaccurate.

Clearer now?

Oh, I forgot to mention - Intel has a Q9550S, which is a 65W TDP version of the Q9550. Do you think it uses 30W less than the regular Q9550?

Re:FAIL

[sexconker]

And your system is unstable.
And your fans are running more.
And your PSU has to work harder to keep those capacitors on your mobo full as they wobble.
And what about your other frequencies?

Measure it from the wall or STFU.

Re:FAIL

[A Friendly Troll]

Thanks for telling me that my system has been unstable for the past two and a half years. Also, thank you for telling me that my two fixed-speed 800rpm fans in the entire case are running "more" now. My PSU, however, is not very happy about learning that it has to keep the motherboard at bay because a CPU using roughly 35W at full load is such a beast, so I can't thank you for that :(

Re:FAIL

[Anonymous Coward]

True, but you can check out comparisons on tomshardware.com and you'll see that the Q9550 has comparable performance and occasionally beats the Phenom II 965, not to mention the fact that the Phenom 2 965 is effectively an overclocked version of the 955.

Add that to the fact that the Core i5 is coming out and performance tests suggest it will outperform with lower power usage and at a lower price, and the Phenom just seems like a bad buy.

Re:FAIL

[man_of_mr_e]

Actually, the Core i7 920 is the sweet spot.. $199 and it blows everything out of the water. X58 Motherboards are more expensive but that is going to change very soon as Intel releases new consumer level i7 chipsets.

Re:FAIL

[Anonymous Coward]

That 60 watt difference is in load usage, not idle. I doubt the GP uses his computer under load for 8 hours per day.

Re:FAIL

[sexconker]

Measure at the wall or STFU.

Re:FAIL

[Anonymous Coward]

I did not have to increase the voltage. In fact, I'm actually running less that the specified voltage.

Re:FAIL

[Anonymous Coward]

Tom's Hardware? you mean that site evidently biased for/by Intel against anything AMD/ATI? Which in some seasons carry just Intel banners? lol

Re:FAIL

[Anonymous Coward]

I'm running less than the stock specified voltage and my aftermarket cooling system is huge thus requiring only slow moving air will little fan power (and no noise). Q9550 overclocks like a beast, that's why I said I could go way above 3.4 Ghz, that's just what I get with under-voltage, dipshit.

Re:FAIL

[PiSkyHi]

Other sites only measured total system consumption from the wall, which is far more inaccurate.

Not according to my electricity bill.

Re:FAIL

[Anonymous Coward]

Get a real argument or head back to 4Chan.

Re:FAIL

[Mozk]

Though power consumption alone turns me off.

You're a circuit breaker?

Re:FAIL

[Anonymous Coward]

Phenom II includes a memory controller on-die, Core 2 doesn't. That could account for some of the difference.

Re:FAIL

[Bu11etmagnet]

Not everybody is good at stealing...

Re:FAIL

[SlovakWakko]

Your CPU uses 95W at its stock speed of 2.8GHZ, not at the overclocked speed of 3.4GHz. If you are willing and able, you could measure the difference and post it - I'd be interested to know the results. I'd expect it to be somewhere near the 150W mark...

Re:FAIL

[Anonymous Coward]

I'd really like to see the justification for how the above can be moderated "troll".

Re:FAIL

[BronsCon]

a 4channer would've said "BITS or GTFO"

Re:FAIL

[Sczi]

Where are you shopping? Pricewatch shows the 920's best price at $275, and Newegg has it at $279. Add in an expensive motherboard and expensive RAM, and you have an expensive system, that 95% of users wouldn't be able to tell the difference. At 130 watts, btw. On the other hand, the 3.0ghz AMD 95 watts at $169, 785g motherboard with infinitely better onboard video, multi-media crap onboard, etc, and you have a killer all around system for like 400 bucks. I do wish AMD had an answer to triple channel ram, though. Intel definitely dominates the high end right now. I'm still kicking around an AMD 939 2.2ghz dual that I wish would get hit by lightning or something so I could muster up an excuse to replace it. I'm on an AMD 2.8 quad (45nm) with 8 gigs of 1066 ram and raid 0, and for most tasks, I can't tell a bit of difference. I suspect upgrading to an Intel 920 would be similar.

What's up with Newegg, btw? They're either getting hit with a dos, or it's really time to upgrade.

Re:FAIL

[Anonymous Cowled]

It was a valid point, but unfortunatley, the /. mods are all on crack.

Re:FAIL

[Anonymous Coward]

Yeah, 2006 called and said that he wants his CPUs back.

They're not even able to compete on power any longer... look at the reviews the 965 draws more power at BOTH idle and load than the core 2 quad 9550S & i7 920 according to the Tom's hardware evaluation.

Add this to they're barely matching performance of the core 2 quad (yesterday's arch), and still being thrashed by i7 with the Tom's Hardware article strongly implying that i5 will also be thrashing Phenom II. AMD will be lucky to get the upper low-end market

I think that Intel should pull the AMD marketing stunt and start labelling the the i7 920 as i7 920 4200+.

So lets see, AMD can't come up with a new arch after 3y, fail at linux GPU support, fail at power utilization, fail at stable air overclocking, etc.

wow sad to see that AMD is now becoming a synonym for fail. I really think that they made a BIG mistake in buying ATI, as they don't appear to have the resources to adequately support any of their businesses even with mfg spinoff. AMD is really beginning to remind me of Apple around the turn of the millenia with powerpc and all of the little speedbumps they kept pushing out while AMD/Intel were going GHz+

OK so I'm a little bitter since I bought a nb with an ATI GPU and found the linux drivers to be absolute crap, and then when queried get sloughed off with well, we really are only supporting workstation customers, but we're releasing the specs, so to get full use out of your shiny new GPU you'll have to wait for OSS driver support as we're likely to legacy your GPU on you before we finish support in the closed drivers, so there! BFD nVidia manages to support more OSes well, and they're a smaller company.

Re:FAIL

[Anonymous Coward]

Microcenter. http://www.microcenter.com/single_product_results.phtml?product_id=0302727

That's hot!

[roger_that]

That's a hot CPU for a hot price. Take that any way you want.

My Computer ...

[Archangel Michael]

.. can now double as still for my homemade vodka

Re:My Computer ...

[sgt scrub]

Brilliant! I was thinking grilled cheese sandwiches but that is WAY better.

Re:My Computer ...

[Inner_Child]

Awesome, now people can rehash their old P4 jokes that weren't funny the first time around!

Re:My Computer ...

[Anonymous Coward]

Oh man! It's not that my jokes aren't funny. It's nobody gets them. Not even me.

AMD...

[mcgrew]

140 watt processor, now with the power of two light bulbs! Space heater no longer needed in the room your computer is in!

Re:AMD...

[squallbsr]

I don't need ANOTHER heater:

I already have:
my Quad Q6600 + 22" monitor
my C2D MacBook + 24" monitor (Dell UltraSharp furnace)
my AM2 X2 + 2x 19" monitors
my P4 3.06 533Mhz FSB (forget which series the chip is from) + 19" monitor

Now if I could only efficiently recirculate air from my 150 sqft office to the rest of the house, I wouldn't need a furnace...

Re:AMD...

[BlackSnake112]

Now if I could only efficiently recirculate air from my 150 sqft office to the rest of the house, I wouldn't need a furnace...

If you have a furnace (or forced air system) see if you can turn on the fan only. The return(s) in the room should pull the heat out of the office to the other rooms. That is if you have a return in the office, a good system has a return (even a small one) in every room. My house built in the mid 1950's has small returns in every room.

Re:AMD...

[MiniMike]

I wouldn't need a furnace...

How would better air circulation eliminate the need for your P4 system?

Re:AMD...

[Anonymous Coward]

No one gives a fuck. Masked arrogance is pure faggotry.

Re:AMD...

[Anonymous Coward]

Butthurt troll...

problematic for some:

[nimbius]

read: if you didnt shit in a marble toilet this morning and start the day trying to figure out which of your sedans to drive to the office, you may find this chips introduction "problematic" from a pricing standpoint.

Re:problematic for some:

[ByOhTek]

With $1000 enthusiast CPUs, the price isn't horrible by comparison.

Still, you can get Intel CPUs that'll run at similar clock speeds for similar prices (except use more electricity), and right now, Intel has the performance/clock ratio advantage over AMD.

So I wouldn't call it /expensive/ in the grand scheme of things, it's just not that impressive either.

Re:problematic for some:

[MBGMorden]

Are you kidding? It wasn't THAT long ago that $245 bought you a budget CPU, and plenty of people without "marble toilets" had computers back then. Yes, it costs a bit more than the $50 budget chips that are available now, but I think you're exaggerating the impact of the price here. It's not that bad . . .

Re:problematic for some:

[PitaBred]

The thing is, it's the value proposition of AMD that's attractive... the motherboards are cheaper, and you can upgrade incrementally with them. You don't have to get a whole new system to upgrade. The AM3 chips fit into the AM2+ sockets, the AM3 chips are compatible with DDR2 RAM as well as DDR3...

Re:problematic for some:

[sgt scrub]

drive? myself? Oh it was supposed to be a +1 funny post. Sorry.

Re:problematic for some:

[Tubal-Cain]

I could, but I can't justify it to myself. Replace a whole PC, and you can do something useful with the old one. Replace a part.... this RAM is still perfectly good, but I can't do anything with it.

Re:problematic for some:

[MBGMorden]

I see your point there, but, for me at least, I eventually run out of "useful" things to do with my old pcs a while back. Router and NAS are handled by dedicated devices (both are running embedded Linux though so I can technically SSH in and still play with em). I've got 3 desktops (Mac, Linux, Windows), a 4th desktop for playing with stuff (currently running Syallable, though will likely switch to Haiku if they get a better installer available), a Windows laptop, and a MythBuntu based HTPC.

Honestly I just don't see many niches for extra computers to do much more for me other than drive up the power bill more. If at all possible I pretty much always go for upgrades, unless a computer is just in need of an overhaul, at which time I still usually salvage SOME stuff from the system it's replacing. And if I get too tired of how my computer looks I'll just buy a new case next time I format it. That makes it seem new, at least :).

Re:problematic for some:

[G2GAlone]

I completely agree with that... AMD is my top choice for clients with simple computer needs. Why? Cheap and easy to upgrade (when said client gets convinced by some meatball relative that they NEED 4gb of RAM in their PC for the 20 minutes of web surfing they do in a day...)

Either way, both AMD and Intel have their markets, I don't think it will be too much of an upset if AMD grabs the midrange market... Hopefully, us consumers can sit back and enjoy as they "competitively price" their latest offerings...

Re:problematic for some:

[TheRaven64]

It was quite a long time ago. The first machine I built myself had a 233MHz Cyrix chip which cost about £50 and plugged into a £40 motherboard. Not sure what the exchange rate was back then, but even if it's the worst it's been for the US dollar it would be about $100 for the CPU. And back then a lot of people I knew didn't have computers. This was over 10 years ago, and back then the most expensive individual component was the hard disk. Since then, CPUs have slowly dropped in price.

Re:problematic for some:

[Anonymous Coward]

I upgrade via whole pc / laptop at a time these days also (just becuase computers are so much cheaper now than they ever have been). But all my old pc's basically turned into a parts heap : ) I'm actually going to cobble together a fileserver from them one of these days.

Re:problematic for some:

[owlstead]

Other people give them away or sell them. Normally I give them away, although I do upgrade as well. Fortunately cases are still pretty easy to get, because I do always want to keep my screw less case.

Re:problematic for some:

[hawkingradiation]

And virtually all systems with AMD processors support virtualization, while with Intel chips the bios must enable it first at boot. No more Sony for me. As another dig against Sony, I went to my local Best Buy to get a video camera and was informed that virtually all cameras support Mac and PC except for the Sony ones which will only work on WIndows.

Re:problematic for some:

[BikeHelmet]

You must be relatively new to computers. If I kept every whole PC I ever made, I'd have a dozen or so sitting around doing nothing.

There's only so many NAS's one can assemble!

Re:problematic for some:

[Tubal-Cain]

It's true I haven't owned many computers, but that's more due to a slow upgrade cycle.

As for only being able to build so many NASs:
Whole computers make much better gifts/donations than parts. I might not have much use for an idle piece of RAM, but the best use any of my family members could come up with would probably involve arts & crafts.

is this for desktop or server ?

[C0vardeAn0nim0]

if it's a server class part, i think they'd do better emulating sun's T2 part. 8 multi-thread cores, a single FPU unit shared by all cores and some logic to improve encryption and networking. this with x86 support would give Xeon a run for it's money.

now, as a desktop part, i think it's idiotic as hell. a low power chip with a decent software stack to offload certain proccessing tasks - like video and audio encoding - to the GPU (wich they also make) would do much better in terms of performance per watt.

but then, none of those options would serve as silicon penile extension, right ? no bragging rights like "dude, i just got a quad-core, 3.4 megahurtz chip! duuude !"

Re:is this for desktop or server ?

[Flea of Pain]

Wait what? This also serves as a penile extension??!!?? I"ll take 3!

Re:is this for desktop or server ?

[sgt scrub]

It better extend it more than 3 times, even at that price!

Re:is this for desktop or server ?

[reub2000]

INASA (I'm not a server admin), but why would you put a power hungry chip in a server? Data centers always seem to be having problems with cooling.

Re:is this for desktop or server ?

[Bill, Shooter of Bul]

In most cases yes, but not always. In any case this chip is meant for desktops. Anyone using a desktop chip for a server has other issues as well.

Actually brings back bad memories of an old company buying the absolute cheapest server possible to save money. A dell with a Celeron(pentium III based) and 256 megs of ram! Yes, it was a freaking celeron in a server. And boy did it suck. My two year old desktop with a Pentium 4 ran rings around it. It was so, so sad. I can only imagine the laughter when they migrated to a new data center with that.

Re:is this for desktop or server ?

[reub2000]

It depends on what you are serving. If your serving a shared folder on a home network, a Celeron is plenty. Hell, it's probably enough to serve a small website with only a handful of visitors.

I'm not sure what's wrong with a desktop chip in a server. Besides for SMP being disabled, what's the difference?

Re:is this for desktop or server ?

[Kz]

... T2 part. 8 multi-thread cores, a single FPU unit shared by all cores...

You're thinking about the T1, the T2 has one FPU per core

Re:is this for desktop or server ?

[Bill, Shooter of Bul]

The problem is the amount of stuff we were serving. The server was pegged during peak hours. Ran at 70% most of the rest of the time. It did a lot of on the fly graphics creation, number crunching and database access with every request. It wasn't exactly google's load but it wasn't just a web server with static content either. Would have been fine with even a lower end xeon and a bit more memory. The price difference was all of $200.

This is midrange?

[Lord Byron II]

How is a $245, quad-core chip considered mid-range?

Re:This is midrange?

[SkankinMonkey]

Have you seen Intel's pricing? Can't get a decent solution from them including a motherboard for under 500 whereas I just recently built a full Phenom II computer for about 400 (including hard drive). I'm not convinced that Intel is really interested in mid-ranged computing.

Re:This is midrange?

[characterZer0]

Midrange for servers?

Re:This is midrange?

[Stenchwarrior]

Because it's between this

And this

Re:This is midrange?

[kieran]

Don't let the quad-core bit fool you; that'll be low-end in a couple of years, no doubt.

Still, the chip in question is definitely at the upper reaches of "mid-range" in my book. I've just picked up the X3 720 model for my home machine, and that was stretching the wallet as far as I was inclined to.

Re:This is midrange?

[afidel]

Because it's between $100 and $500 which is probably the high end for most PC class processors. Intel has the Core2Duo at the low end, the Core2Quad and the low end Core i7 in the midrange and the faster Core i7 at the high end with a few enthusiast offering at the extremely high end (~$1000). That's basically been the market as long as I can remember which dates back to the early 90's.

Re:This is midrange?

[ByOhTek]

$200-$400 seems to be Intels "mid range", so they are probably comparing it to that.

Re:This is midrange?

[Anonymous Coward]

Awww crap@! My pentium is now the best example of the worst processor? WTF? Its a 32 bit *protected* mode powerhouse! You wouldn't believe the frame rates I get with Doom!

Couldn't you have at least set the bar a little lower. To, I don't know something like this

Re:This is midrange?

[beelsebob]

You suck at pricing up intel kit, you can get a Core2Quad Q9550 and mobo for about $300, or an i7 920 for $350.

Re:This is midrange?

[Albanach]

Don't let the quad-core bit fool you; that'll be low-end in a couple of years, no doubt.

We've been waiting a decade for improvements in multi-threaded processing to take advantage of multiple cores.

Are you suggesting programmers are going to make the dramatic developments in the next couple of years that they have been unable to in the last ten?

Certainly I can see the number of cores increasing at the server end - it's straightforward enough to run one process per client. I'm unsure what's going to change on user desktops that will drive any massive increase in core numbers. Still, I'm prepared to be surprised.

Re:This is midrange?

[Jeff Carr]

I just built a new computer for my parents with an Intel boxdg41ty, E6300, 4 gigs of Patriot DDR2, a 1 terabyte Seagate drive, and a Corsair 400 watt power supply for about 275 after rebates and shipping. It isn't a gaming machine, but it works beautifully for just about anything the average non gamer will throw at it.

Re:This is midrange?

[nobodylocalhost]

huh? what planet do you live on to pay 350 for i7 920? i paid $199 for mine, given it was on a sale, but normal price for 920 are in the low 200 usd range.

here is a link:
http://www.microcenter.com/single_product_results.phtml?product_id=0302727

Re:This is midrange?

[anagama]

I think he was including the motherboard in the price.

Re:This is midrange?

[Mr. DOS]

Eh? I just paid CA$310 for an i7 920 and around the same for an LGA 1366 motherboard. I'd love to know what fly-by-night retailer you're getting stuff from...

      --- Mr. DOS

Re:This is midrange?

[PitaBred]

You still need DDR3 RAM, and an i7 mobo. Those are both more expensive than an AM3 motherboard with DDR2 RAM. Not as fast overall, but it's a fair bit better value for the buck. Not to mention that you can upgrade incrementally with AMD... the AM3 CPU's work with both DDR2/AM2+ motherboards and DDR2/AM3 and DDR3/AM3 motherboards.

Re:This is midrange?

[PitaBred]

You can get an X4 955 for $200 on sale. It's a 3.2GHz/125W quad-core. Not too much wallet stretching, especially compared to this new one.

Re:This is midrange?

[confused one]

Because the high-end is the 8-core workstation sitting next to me at my desk.

Re:This is midrange?

[beelsebob]

Yes, and then you need to buy a $150 mobo, he did after all state he was looking for mobo *and* CPU.

Re:This is midrange?

[beelsebob]

Yes, but if you're going for value for your buck, the Core2Quad Q9650 is as fast as this, uses that nice cheep DDR2 RAM, is cheeper than this AMD chip, and the mobos are nice and cheep too. So all together, AMD's looking pretty crappy atm.

Re:This is midrange?

[k_187]

Honestly, I think its been a chicken and egg problem for a while. Nobody used multiple cores on the desktop because software didn't take advantage of it and the software didn't take advantage of it because nobody had multiple processors. Intel and AMD realized that if they wanted to keep on the MOAR POWAR treadmill, they were going to have to start packing more cores into the processors. There's no reason to think that programmers won't catch up eventually. We're already seeing stuff take advantage of dual cores.

Re:This is midrange?

[san]

Multicore CPUs are more about what CPU designers can deliver than about what people can actually use.

Because the limiting factors in single-cpu performance have been memory latency and instruction-level parallelism for the last half decade, there has been very little progress in single-core cpu performance over that period of time.

Both these problems won't find a solution any time soon, so don't expect the cores of your 64-core CPU of 2015 to be much faster than the cores of today.

Re:This is midrange?

[dwinks616]

http://dictionary.reference.com/browse/cheap ... ...

Re:This is midrange?

[GooberToo]

Are you suggesting programmers are going to make the dramatic developments in the next couple of years that they have been unable to in the last ten?

Yes. Ten years ago SMP was not common. These days it is. Don't let complexity vs performance gain confuse your sense of direction. Without a doubt, there is an ever increasing push to take advantage of multiple cores, where it makes sense to do so. Already more and more games attempt to break audio, graphics, physics, loading/caching into multiple threads of execution with varying degrees of success.

Additionally, applications like ray tracers, cad, modelers, image manipulation, etc...continue to increase their concurrency. Hell, even python, a classic lead foot in the language world, ever trying to remain anchored to the ground, has continued to be pushed into better SMP/concurrency support. Java continues to add concurrency support in its contains and VM.

So yes, absolutely, things are charging. No ifs, ands, or butts. It just may not change as rapidly as we'd all like.

Re:This is midrange?

[Fulcrum of Evil]

they can be slower for all I care, so long as I can turn them on and off quickly to match load - 1.8G parts that burn 20W and do everything I want suit me just fine.

Re:This is midrange?

[Anonymous Coward]

We've been waiting a decade for improvements in multi-threaded processing to take advantage of multiple cores.

Maybe you've been waiting. I upgraded from MS-DOS to a multitasking operating system.

You don't need multithreading to reap benefits from multiprocessing. My experience is that if your box is busy enough that you want to speed it up, your load is probably already parallel.

Re:This is midrange?

[Ant P.]

I'm unsure what's going to change on user desktops that will drive any massive increase in core numbers. Still, I'm prepared to be surprised.

Browsers. Chrome already does multi-process, Firefox 3.5 uses multithreaded audio/video, and Javascript has explicit (albeit uselessly heavily sandboxed) threading support now as well.

Re:This is midrange?

[mcvos]

Where do you shop? There are much cheaper shops out there. Anything over $300 for CPU+MB isn't midprice anymore.

Re:This is midrange?

[Anonymous Coward]

Cheepy cheepy cheep cheep

Re:This is midrange?

[louiswins]

I don't know if this was on purpose, but I'm tempted to buy that Pentium I just because it costs $13.37.

Re:This is midrange?

[SBrach]

I have had really good luck with intel boards and you can pick up a decent one for $60 bucks. Add any one of these cpus for $149-$349 plus $45 for 4GB of Ram and you have a pretty fast rig for $254 - $454.

Re:This is midrange?

[amorsen]

Both these problems won't find a solution any time soon, so don't expect the cores of your 64-core CPU of 2015 to be much faster than the cores of today.

At least they'll be connected to more cache and faster memory. Hopefully the third level cache will be shared among several cores.

Re:This is midrange?

[bitrex]

the AM3 CPU's work with both DDR2/AM2+ motherboards and DDR2/AM3 and DDR3/AM3 motherboards.

The AM3 CPUs only work with AM2+ motherboards if your motherboard vendor decides to provide a BIOS update to support them, and certain motherboard manufacturers (cough ASUS) have been reluctant to provide BIOS updates to upgrade "obsolete products" for obvious reasons.

Re:This is midrange?

[san]

That's true, and it will make a big difference for some applications, but once the problem set fits in cache (and for many applications it already does) you won't get any faster.

The truly interesting thing, though, is that CPUs will look more like GPUs when there's many of them. My prediction would be that the distinction between cores will get fuzzier and CPUs and GPUs will be integrated into devices that can reconfigure themselves into a massive number of very simple cores (like a GPU), or a smaller number of superscalar traditional CPU cores.

The relative complexity of instruction decoding will get smaller and smaller as the number of transistors increases, so reconfiguring CPUs like that will get cheaper and cheaper.

Re:This is midrange?

[man_of_mr_e]

$199 is pretty much Microcenters everyday price.. they treat it like a sale, but it's been that way for well over a month and shows no sign of changing.

Re:This is midrange?

[man_of_mr_e]

The motherboard is a little more expensive, but DDR3 isn't.. you can get 6GB matched sets for $99. And Intel is coming out with consumer i7 chipsets very soon which will bring down the motherboard price significantly.

Re:This is midrange?

[dangitman]

Well, Snow Leopard is coming out next month, and one of its goals is to advance multi-threading at the OS level.

Re:This is midrange?

[bemymonkey]

Actually, the only way I can really bog down the system these days (not including stuff like X264 encoding, obviously) is with I/O. When 3 programs are trying to write to the drive simultaneously, everything slows to a crawl...

SSDs are a much more welcome development than more cores in CPUs, IMO.

Re:This is midrange?

[BikeHelmet]

But then you lose the upgrade path. If you need a RAM upgrade, you're locked into DDR2, while DDR3 could very well be significantly cheaper when you finally need more RAM.

Re:This is midrange?

[Zxern]

only if you live within a reasonable distance to a microcenter store. That price is brick and mortar only. The price every where else is about $280.

Re:This is midrange?

[FrankieBaby1986]

More Cores, More Applications. Hell, one core to handle desktop compositing stuff with the GPU's help, one core to crunch the video I'm watching on youtube, another core to Rip a CD to mp3, and other to deal with my IM app, my browser, my file browser, and all the other little tray apps and background tasks that wake up once in a while. Bring on the cores.

As non-hardware geek, why would I want this?

[Anonymous Coward]

Someone please explain

Re:As non-hardware geek, why would I want this?

[azior]

the ladies will love you for it, if you know what i'm saying...

Re:As non-hardware geek, why would I want this?

[avandesande]

What are you doing here?

2 things

[ILuvRamen]

okay, 2 things about this: If their brand new, top of the line processor is only shooting to take out the midrange market because it's still worse than the best intel ones, that's not a real good sign. Also, you might not be overly concerned about power comsumption but in a small apartment or bedroom or home office, that room is going to get really hot really fast. Combine it with my 8800GTS and you've got a nice little space heater, which is great in the winter but not so much in the summer.

Robust support

[Anonymous Coward]

and their motherboard support for this new 140 watt processor is robust."

Given the sort of heat sink a 140W CPU will require, the motherboard supports had damn well better be robust.

Q6600

[Hadlock]

I have my sights set on the intel Q6600 (2.6(?)ghz quad core2)... as soon as the price dips below $175. It's been the same price since Dec 2007, but it overclocks to 3.4ghz on the stock cooler and dissipates only 95w, meaning it plugs into pretty much any C2Duo motherboard. For now though, my 2.4ghz core 2 duo is fast/powerful enough to do anything, including run folding@home while running TF2 + 20 windows of firefox with 5-10 tabs each.

Re:Q6600

[CrimsonKnight13]

You can easily hit 3.6GHz on air w/ an FSB of 1800 (mind you, I use an aftermarket cooler - Thermalright Ultra 120 Extreme). It's a great chip to overclock with & I'd recommend it for anyone that wants to build a great PC with a lower budget.

Re:Q6600

[Anonymous Coward]

the q6600 is out of production. the q9550 is faster, uses less power, and runs cooler than a q6600 and this amd phenom. i run mine at 3.4ghz with 1.15V. thats less than stock voltage.

Re:Q6600

[Hadlock]

Does the q9550 do virtualization though? The q6600 does.

Re:Q6600

[iamhassi]

I've had a Core 2 Duo E4300 running at 3.0 ghz for almost three years now and I haven't found a reason to upgrade. Friends with quad core report no increase in speed or performance, and the only thing that would encourage me to upgrade is a more smp friendly OS but the offerings from Microsoft (Vista and Windows 7) have been pretty poor lately.

So are we done with the mhz battle? Is ~3ghz the breaking point? We've had Xeon 3.0GHz cpus for over 5 years now. That's a long time to not see a jump in speed, what happened to "doubling every 18 months"? We should be around 24ghz by now.

Re:Q6600

[Rudeboy777]

Sorry, that CPU will never decrease in price. It will be phased out while remaining at its current price. I recommend combing EBay and Craigslist for a used one if you are convinced this is what you want.

Re:Q6600

[pinkishpunk]

unlike intel amd does not segment their cpus into a lot of different subtype that lacks features, only the sempron types lacks virtualization support, the x2,x3,x4 types all has it.

Re:Q6600

[arazor]

Yes the Intel Core2Quad Q9550 supports virtualization.

Re:Q6600

[Reapman]

Agreed.. I remember winning a 3.2 Extreme Edition CPU 4-5 years back or something like that. My 2 year old C2D I've been able to overclock to 3.2, but the Mhz battle is truely dead, at least for now.

Not that it's a bad thing, CPU's are way more efficient and still faster then the CPU's 5 years ago, but it was fun watching the battle. AMD isn't much more then a whimper right now.

Re:Q6600

[Anonymous Coward]

yes it does

Re:Q6600

[PotatoFarmer]

That's a long time to not see a jump in speed, what happened to "doubling every 18 months"? We should be around 24ghz by now.

If you're referring to Moore's law, it's not a doubling of speed every 18 months, it's a doubling of transistor counts. Clock speed has never been part of that equation, no matter what intel's late-1990s marketing department would have you believe.

Re:Q6600

[Anonymous Coward]

Yes, only the 8xxx series don't do virtualization, AFAIK. You might double check the spec sheet to verify.

The bigger problem for me is that the 8000/9000 series use a 1333 FSB, which means you've only got one step up for overclocking rather than 2 :(

Reason I haven't bought a chip above the E5400 yet, best clock multiplier means best overclock potential for the same cores.

Pentium E6300 is worth a looksee though, Dual Core, second best clock multiplier behind the E5xxx series, supports VT, and is only 80-90 bucks :)

Honestly for the cost of a FSB based Quad core now you can get a QPI Quad Core with hyperthreading. And assuming you're not running cache hungry apps, it's got plenty of memory bandwidth to saturate all '8' threads.

Re:Q6600

[NervousNerd]

Actually the new Sempron 140 has AMD-V.. It can also unlock into a dual core. As always, even though AMD can't compete on the high end, they can almost always compete on the low end and lower mid-range.

Re:Q6600

[MarkvW]

Don't stop him. He's on a roll.

Re:Q6600

[mdm-adph]

The newer batches of the Q6600 (2.4 GHz stock, by the way) are terrible overclockers -- you'll have to buy one of the older ones off of Ebay to have any chance of getting it above 3.2 GHz (even experienced overclockers had tons of trouble with the newest Q6600's).

Just get a Q9400 -- it's both faster, cooler, and has more overclocking headroom.

Re:Q6600

[Ant P.]

Why would you need 200 webpages open at once while you're playing a fullscreen game?

Re:Q6600

[Hadlock]

If you really have to ask that question, you really shouldn't be posting on slashdot. But I'll humor you :)
 
TF2 has what's called windowed noborder mode which allow you to alt+tab between it and other apps with zero impact on gaming performance on modern computers. I play arena on tf2 (shamless plug steamcommunity.com/groups/lemuria) which has a lot of sit out time, similar to counterstrike. Its convenient to alt tab between tf2 and firefox. Also, with a 4 screen setup on a computer that is never turned off and has a 3+month avg uptime, there's no reason to ever close a window if you have an efficent windowing solution. The only time firefox crashes is usually a day or two after it automatically downloads and applies an update in the background and you decline to restart. So yeah, 200 pages, between forum threads, slashdot, guitar tabs and more isn't unreasonable. I guess you could call me a power user:) Now queue the posts telling me a more efficent way to do it that true power users use.

Re:Q6600

[mdm-adph]

False, Intel kept this marketing strategy going well into the 2000's. -_-'

Re:Q6600

[mcvos]

Why would you need 200 webpages open at once while you're playing a fullscreen game?

Because I can.

Re:Q6600

[nog_lorp]

Of course it does. (... double checks)

As far as I know, AMD and Intel aren't even building new chips without virtualization. Their goal is to make it ubiquitous so it can actually become a consumer technology (for virtual pc stuff, anti-virus vendors etc).

Re:Q6600

[Anonymous Coward]

TF2? pfff I can do the same with Quake2 but no I don't have 200 tabs open, I know what bookmarks + tags are for ;).. could you please think of the false positives you're giving to Google analytics?

Re:Q6600

[MogNuts]

That's not being a power user, that's just being silly. Don't u worry about having sensitive data stored in cache for other websites to view (simply clicking "sign out" doesn't do everything vs. exiting) or just to exit firefox once in a while to reclaim memory or to clean up garbage in memory?

Re:Q6600

[Anonymous Coward]

... actually, Moore's law is not about density either. It is just an observation in the cost per transistor. As in, the nominal cost for a transistor halves every 18 months.

The era of the silicon chip is gone.

[For a Free Internet]

AMD and Intel are just running on its fumes. Silicon (Si) is inherently limited by its inorganic composition which means it produces lots of heat especially when it is on the Web. All the smart engineers at the secret R&D labs are working on organic computing: solving the paradox of user interface versus wattage by harnessing the power of bacteria to create a new paradigm of information that is multi-dimensional. Instead of "processes" and "treads" and "HTML" we will have gases and sugars dancing to the rhythm of our wildest imaginations. And one more thing... you will not need your eyes any more since the two-dimensional "screen" and "paper" metaphor will be replaced by a revolutionary direct access to pure consciousness. Buy my book.

Re:The era of the silicon chip is gone.

[sgt scrub]

Buy your book or smoke your weed? If that later produces that output I'm buy'n!

Re:The era of the silicon chip is gone.

[Anonymous Coward]

mmm... fumes...

Re:The era of the silicon chip is gone.

[vivek7006]

Hey, is this you Deepak Chopra? Now you have started trolling on Slashdot. Recession must have hit you badly I guess

Re:The era of the silicon chip is gone.

[Jon Abbott]

This reminds me of that one scene from True Stories...

Re:The era of the silicon chip is gone.

[Anonymous Coward]

I'd rather buy some of what you've been smoking.

Re:The era of the silicon chip is gone.

[Anonymous Coward]

Can I buy some pot from you?

More cores?

[DoofusOfDeath]

Anyone know when we should expect the first CPUs from either Intel or AMD that have more than 6 cores?

Re:More cores?

[lukas84]

Intel already has the 6 core Dunnington CPUs out since over a year now. The new nehalem based 6 cores are due Q4.

Re:More cores?

[DoofusOfDeath]

Intel already has the 6 core Dunnington CPUs out since over a year now. The new nehalem based 6 cores are due Q4.

Thanks, but I'm specifically asking about > 6 cores.

I'm doing some work regarding scalability of certain algorithms on multi-core systems, and if a new CPU with even more cores is just around the corner, I might delay a hardware purchase until then.

Re:More cores?

[lukas84]

Beckton, the 8 core / 16 threads Nehalem CPU will be out in Q1 2010.

http://www.theinquirer.net/inquirer/opinion/1050976/intel-bunch-fun-cpus-moves-2010

Re:More cores?

[Anonymous Coward]

AMD has 6 core Istanbuls in production now, and will have a 2 die 12 core MCM version Real Soon Now.

Re:More cores?

[DoofusOfDeath]

Thanks.

Re:More cores?

[DoofusOfDeath]

AMD has 6 core Istanbuls in production now, and will have a 2 die 12 core MCM version Real Soon Now.

Do you mean a total of 12 cores in a single chip package?

Re:More cores?

[kieran]

Talk to Intel - if you're doing something that might help them sell multi-core chips, they might bung you a pre-release chip for free/cheap.

Failing that, is using more cores per die very different to using multiple processors?

Re:More cores?

[lukas84]

MCM = Multi Chip Module

Re:More cores?

[networkBoy]

no.
in a multi chip module (MCM), intel calls it a Multi Chip Package (MCP).
It's two single chips of 6 cores each in one package.
It is two chips in one socket.
-nB

Re:More cores?

[DoofusOfDeath]

Failing that, is using more cores per die very different to using multiple processors?

Price-wise, I'd say yes. Motherboards with > 2 CPU sockets tend to be super-expensive.

I believe it's a lot cheaper to get 8 cores using 2 chips than it is using 4 chips, for example.

i7 920 130watt - $280, x4 965 140 watt - $245.

[millisa]

When it comes down to processor comparisons, I see very little compelling about this new AMD proc. The i7 920 is going to outperform it at most things, uses less power and is only 35 bucks more. Eventually for those of us always-on users, even the 10 watt savings of the i7 is going to kill the slight price advantage.

The only thing I see interesting here is the fact that you have more commodity boards to choose from, could do a slower upgrade (re-use your ddr2!) but this isn't any different than the currently line of quad proc amd chips, many of which can be had for cheaper and use less power.

Come on, AMD, you can do better.

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[lukas84]

Yeah, but the i7-920 requires a mainboard with triple channel memory and a quick path interface. They're more expensive than the AMD board, making the price comparison a bit more difficult.

That said, i bought an i7-920, i think it's the better choice - so far, i haven't been disappointed. We have a few new servers with 5540 Xeons, and they're absurdly fast.

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[beelsebob]

You can get an i7 920 for $200, so not only is it faster and lower power, it's also cheeper.

http://www.microcenter.com/single_product_results.phtml?product_id=0302727

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[Bill, Shooter of Bul]

Its my understanding that with cool & quiet, the cpu scales down its Mhz and power consumption when not busy. Last Tom's hardware article I saw comparing power usage showed that the AMD's used significantly less power when idle, despite their higher wattage ratting. So if you're leaving it always on, you shouldn't notice much of an energy difference, unless its crunching away @ 100% all day.

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[Hadlock]

Yeah, but you have to pay the "i7 tax" for a new $220+ motherboard. Core2Duo/Quads will happily plug into any $50 motherboard you (already) have.

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[elrous0]

any significant processor upgrade is going to mean a new mobo too. Unless you're upgrading your processor every few months, it's almost always stupid to keep a mobo that's probably just as obsolete as your old processor (if the new processor will even work with it at all) just to save the relatively trivial cost of a new mobo.

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[beelsebob]

You can get hold of S1366 mobos for about $140 these days. Still pretty pricey, but no where near as bad as they were recently.

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[StuffMaster]

No I can't. You forgot to include gas (since there's no shipping).

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[Hadlock]

I'm not able to look at that board's specs right now, but most people I know skip the bargain bin motherboards and go for Asus brand (I prefer genuine intel myself) which are generally mid to high priced boards. Core2Duo boards of that caliber are still $100-120

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[Prien715]

According to your link, that proc only runs at 2.67 Ghz, which is a far cry from 3.4.

Maybe the link is wrong?

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[beelsebob]

So? The i7 is a massively better design than the Phenom II. The i7 920 benchmarks faster than any of the Phenom IIs by a large margin.

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[man_of_mr_e]

Megahertz mean nothing. The i7 920 is almost 2x faster than the 965 at various benchmarks.

Re:i7 920 130watt - $280, x4 965 140 watt - $245.

[Anonymous Coward]

Yea if you live in the middle of nowhere near one of their brick and motar stores. In store pick up only

The low platform cost makes it a winner

[Anonymous Coward]

The $245 CPU price does not tell the whole story: you should also factor in relatively low cost of the motherboard and memory options relative to the sadly spendy i7 platform.

Re:The low platform cost makes it a winner

[beelsebob]

Yes, but the chip doesn't compete with the i7 at all. It competes with the Core2Quad Q9650, which has plenty of cheep boards out there.

Re:The low platform cost makes it a winner

[Anonymous Coward]

Yes, but the chip doesn't compete with the i7 at all. It competes with the Core2Quad Q9650, which has plenty of cheep [sic] boards out there.

Why would I want a board that sounds like a baby bird?

Maybe they make that strange noise because they're so cheap. (^_^)

BTW, Beelsebob, given your spelling, do you ever worry that you sold your soul to Santa? :-)

News?

[Anonymous Coward]

I don't see how this is news... It's the same technology on the other Phenom IIs, except with an overclocked multiplier and price.

Re:News?

[TheKidWho]

It's nothing special, just a higher binned Phenom II.

Considering you can get an i7 920 for $199 at microcenter, this seems kinda useless since the i7 will overclock better and run much faster.

Space heaters

[Anonymous Coward]

That is why we have our computers in the cold basement, heat down there is good year round.

The re-birth of the BTX form factor

[kriston]

Finally, the BTX form factor will be reborn. So long have we been complacent with the temporary trend of cooler running chips there is finally the need for the BTX form factor to extend beyond the Dell desktop.

Re:The re-birth of the BTX form factor

[Rudeboy777]

It'll be a true rebirth - Dell hasn't used BTX motherboards for quite some time now. I'm not holding my breath however.

Re:The re-birth of the BTX form factor

[BlackSnake112]

Then all of these Dell GX745, GX755, GX960 machines with the video card fan on top of the card must be all smoke and mirrors. Dell is using only BTX style in their tower desktops. The CPU is in the front and center (mostly front center) on the motherboard with a big fan pulling air in. The motherboard is installed on the other side of the case.

Re:The re-birth of the BTX form factor

[kriston]

That's interesting because all of my Dells are BTX. I guess they're old now.

Re:The re-birth of the BTX form factor

[toddestan]

I've seen some new HP's that appear to be using BTX too. I say 'appear' because I haven't checked closely to see if they really follow the standard, or are just some propriety thing that adopted many of the BTX characteristics.

I'm getting old

[stangbat]

When I put my last machine together in September '08 I went with a 65 watt X2 5200+ (Brisbane). Cool and quiet won out over hot and noisy. Or perhaps you could say slow and boring won out over fast and exciting. Hmmm, maybe there are deeper personal issues at root in this decision...where did my comfortable shoes go?

Re:I'm getting old

[maxume]

Nah, CPUs got fast. It went from, say circa 2000, where lots of tasks would use most of the processing power available on a cheap computer to circa 2005 where some small fraction of tasks (and poorly written flash) use most of the processing power available on a cheap computer. For the majority of users, everything since then has been a numbers game, with little real impact.

Re:I'm getting old

[anagama]

Besides the quality of life features you get with a quiet lower power chip, there are ongoing cost savings. You're saving 75 watts per hour, 1800/day, and 657,000/yr over the 140 watts for this chip. Assuming 10c per kW hour, that's $65.70 per year in electricity costs you aren't paying. If you run your system for three years, you save $197.10.

NOTE: I'm assuming the wattage figures for both chips are the peak levels of course, although idle speeds for both are likely lower and the difference between power usage may be more or less.

Re:I'm getting old

[TheKidWho]

Sort of.

For high resolution gaming, ie with SLI video cards @1920x1200 or 2560x1600 you really need the latest and greatest, a 3.6-4.0ghz overclocked CPU is a must or else it bottlenecks the SLI Videocards.

For people doing HD Video/Audio work, having a fast processor is also necessary.

For anyone doing Scientific/Engineering computing, faster is always better as you can never get too fast the simulation will just get finer.

For anyone doing 3D Rendering, again faster is always better because the faster your processor is, the more effects/higher resolution you can run at.

For the average home user, they probably don't need anything better than a 2.0ghz Core 2 Duo.

Re:I'm getting old

[clem]

Check behind the footrest of your La-Z-Boy recliner.

Re:I'm getting old

[maxume]

Hence the hedging and waffling; lots of...small fraction...majority of.

If you take face time (which I admit is overwhelmingly favorable to my argument compared to cycles used/cycles available), computers spend far more time waiting for users than vice versa (because of the huge numbers of people that check email and the like).

Re:I'm getting old

[flatulus]

"You're saving 75 watts per hour, 1800/day, and 657,000/yr over the 140 watts for this chip."

I know what you're trying to say but you're saying it wrong.

"75 watts per hour" is a nonsensical quantity. 75 watts is 75 coulombs per second. Multiply 75 coulombs /second by 3600 seconds and you get 270,000 coulombs. But that is just 75 watt-hours, which is what you were actually trying to say. So, in 24 hours, at 75 watt-hours/hour, that is 1800 watt-hours/day (which is still a rate of 75 watts).

The watt is a "rate of charge per unit time" measurement. Saying you consumed 1800 watts in a day makes no sense. It is 75 watts * 24 hours giving 1800 watt-hours.

I see this mistake made so often I wonder why we even bother trying to teach math anymore (sigh).

hard comparison to make

[buddyglass]

• AMD Phenom II X4 965 - $249
• Intel Core2 Quad Q9550 - $219
• Intel Core i7 920 - $279

Article shows that performance is roughly equivalent beween the Q9550 and Phenom 965, with the AMD part enjoying a slight advantage if you look at all the benchmarks together. This while costing $30 more and consuming more power.

Would be interesting to see a comparison of the i7 920 with the Phenom. I'm guessing the 920 would outperform, which is what you'd expect since you're paying $30 more.

Additional Coverage, Benchmarks, Power #s at HH

[MojoKid]

HotHardware also covers the new chip here and full system power is tested in addition to benchmark numbers: http://hothardware.com/Articles/AMD-Phenom-II-965-Black-Edition-CPU-Review/

38 C ain't that hot

[Kyont]

I recently completed a home-build with this very CPU. Sure, the chip was the single most expensive piece, but with NewEgg combo deals and shipping discounts, I got the entire machine for about $600, including a smashing new case, plenty of RAM and disk space, extra USB ports and two disc burners. That's mid-range in my book.

I'm sure some of you hardware nerds will smack me down for one reason or another, but as a starting point I just installed the AMD factory CPU cooler it came with. I don't know what wattage it's pulling, but the CPU temperature is holding very steady at about 38 Celcius, and the fans don't even seem to be working very hard for that. It's working great, and at those temperatures, it should do fine for years to come.

My $0.02 on the very rare occasion of having first-hand experience with the actual hardware in the story.

Re:38 C ain't that hot

[JumpDrive]

I'd mod you up.
But this is the customer line they are looking for. We currently use 6 computers with Phenom II 940's onboard at this time for data analysis. When comparing these with Intel computers the cost would have been much higher. What we found repeatedly during matrix calculations was that the performance vs cost just wasn't a contest.
I am really really beginning to wonder how much calculation stress people put on their computer in this mid-range market, because I hear and see all of these benchmarks and people discussing their performance and I just don't see a cost performance comparison. I have talked with other people who do some of the same type of work and they see the same thing.
We do take some added cautions with air circulation just to make sure that we don't run into issues. We repeatedly have cranked these things for 24 hours straight with cpu's running at 70 to 95% of capacity and actually had one occasion where the systems ran over a weekend (limits weren't set correctly).

Re:38 C ain't that hot

[Pentium100]

At that temperature it's cooler than my dual Opteron 270 setup which runs at ~60 degrees on full load (when the room is at 21 degrees).

Re:38 C ain't that hot

[Nuno Sa]

Yeah. Intel's TDP definition is different from AMD's. AMD reports the maximum wattage the CPU can burn at full load. Intel reports a "typical" number.

If this was an Intel CPU the number would be something like 90W or 95W for the very same CPU.

Anyway, everybody knows that. You can't trust vendor's numbers. Just do your own checking with a power meter.

Re:38 C ain't that hot

[mcvos]

I don't know what wattage it's pulling, but the CPU temperature is holding very steady at about 38 Celcius, and the fans don't even seem to be working very hard for that. It's working great, and at those temperatures, it should do fine for years to come.

What wattage it's pulling is highly relevant, however. That's the amount of energy turned into heat, after all. Did you measure this while idling, during typical use, or at full load? 38 degrees while not doing anything special is not anything special. If you can keep a 140W processor at 38 degrees at full load, that'd be quite spectacular.

Re:38 C ain't that hot

[hattig]

Well AMD have started using an ACP number now that is a bit closer to what Intel's TDP definition was.

At a platform level it appears that the AMD solution is still more than competitive on power consumption with Intel's similar offerings (according to the review at Tech Report anyway), so I wouldn't worry too much about the 140W figure. Especially if you're going to add a fast graphics card anyway, which will eat even more.

Re:38 C ain't that hot

[Retric]

The cost benefit curve in the consumer CPU space is strange. A 10% faster might be worth twice as much or be totally useless.

EX: Real time playback of HD movies. If your CPU can't do that it's a pain. However, if it can then extra speed is pointless for that task.

Re:38 C ain't that hot

[BobZee1]

Your $0.02 is not rare because you didn't make a single point. You don't know what wattage you are pulling? Why spout irrelevant numbers? Watts eventually equally heat. good job with the 38C - I wish my processor ran that cool.

Re:38 C ain't that hot

[hairyfeet]

Well I can say that I am moving my AMD 7550 boxes nearly as fast as I can get them together, in fact I have a customer waiting on the one whose parts haven't even arrived here yet. All I get from my customers is them going on and one about how fast and nice and quiet they are. I'm guessing for a good 90% of the average Joes out there the price difference for the Intel parts just isn't worth it.

And I liked how quiet the last AMD 7550 was, even when running stress tests, that I went ahead and built one for me. With 8Gb of RAM, dual burners, a 4650 1Gb GPU, and XP X64 I spent maybe $600, not counting the $100 or so I'll get back in rebates. This baby plays 1080p smooth as butter and is quiet as a church mouse while doing it. You really can't beat the "bang for the buck" that AMD has now, and this is from a life long Intel man.

Re:38 C ain't that hot

[man_of_mr_e]

I really don't understand why anyone would buy the Phenom.. At $245 it's $46 more expensive than the Core i7 920 and performs significantly worse. The 965 isn't listed there, but 955 is, and it's passmark rating is 3,571 while the i7 920 is rated at 5,440. And that's not even considering the fact that you are using triple channel memory access versus dual channel, etc..

Granted, you can get AM3 motherboards cheaper than X58 boards, but Intel is coming out with more consumer i7 chipsets very soon.

Re:38 C ain't that hot

[BikeHelmet]

I just picked up an Athlon II X2 245, to give a boost to my aging Athlon X2 rig.

I've overclocked it to 3.55ghz, and I must say I'm impressed with the power usage.

Full specs:
Athlon II X2 245 2.9ghz @ 3.55ghz (65w, 45nm)
2x2048MB Corsair XMS 800mhz DDR2 @ 920mhz
Asus 8800GS (got it on sale for $40 CAD one year ago!)
Asus M3N78-PRO
Corsair HX620
WD Caviar 640GB

Now, according to my Kill-A-Watt, my draw from the wall is...

WinXP Desktop - 95 watts.
Left4Dead - 120 watts.
MediaCoder - 100 watts, or 130 watts

With my Athlon X2 4000+ (2.1ghz @ 2.65ghz), and no other hardware changed, my draw from the wall was...

WinXP Desktop - 125 watts.
Left4Dead - 160 watts.
MediaCoder - 155 watts.

Note: For MediaCoder I was encoding H.264-mp3 on both cores, to an mp4 container, for upload to Youtube. For my Athlon II, it seems to alternate between 100 watts and 130 watts, depending on the complexity of the video. First pass seems to be mostly 100 watts, while second pass is 130 watts the entire encode.

I'm impressed by AMD's wattage ratings. They're way better than last generation, and they sure don't compare to Intel's!

Re:38 C ain't that hot

[ozbird]

... including a smashing new case ...

Pete? Is that you?

Re:38 C ain't that hot

[JumperCable]

Show us your part list.

Re:38 C ain't that hot

[Nuno Sa]

Yes, I agree 100%.

However everybody still compares AMD's maximum with Intel's "average" numbers directly.

In this instance AMD has a better moral ground, I suppose. I still like Intel and AMD both :-)

Re:38 C ain't that hot

[Anonymous Coward]

Exactly for the reason you stated. Existing vs. non-existing product. Very soon doesn't exist until that very soon time arrives. Additionally, those are just benchmark numbers. The average person won't notice the performance difference so cheaper is better.

Re:38 C ain't that hot

[Penguinoflight]

Intel will be coming out with new i7 chipsets, they just won't be able to work with any of the socket 1366 processors.
It should be noted that Microcenter does not ship at those prices, nobody does. If you're lucky enough to have a MC in driving distance good for you, otherwise you'll be stuck paying a LOT more for an i7.

The 965 isn't even on that list, but you shouldn't really trust a single benchmark or source to determine the advantages of a solution. Even the 940 and 955 beat the i7 920 in significant benchmarks, like winrar - even though it's heavily multithreaded.

The real advantage to the phenom II's is the ability to use ddr2 or ddr3 memory; ddr3 is very expensive in clock speeds that's actually faster than ddr2. Between spending half as much for a motherboard and half as much for the ram you come out significantly lower with AMD.

Re:38 C ain't that hot

[mgblst]

Well I can say that I am moving my AMD 7550 boxes nearly as fast as I can get them together

Ok so this could mean you are moving 100s a day, or one a week.

Re:38 C ain't that hot

[man_of_mr_e]

But it's not cheaper. It's about the same price. If you pay $245 for a processor and $100 for motherboard it's the same as paying $199 for the processor and $150 for the motherboard.

The average person who won't notice a speed difference isn't going to be buying $250 processors anyways, they'll be buying $50 processors and $85 motherboards.

Re:38 C ain't that hot

[metaforest]

I'm running a Q6600 @ 3.65GHz with 8GB of 1066MHz running an Ubuntu LAMP server (Headless)

It cost me less for the CPU 2 years ago than this AMD costs today. The Q6600 runs faster, if I go by the synthetic benchmarks. While I like rooting for the underdog, AMD is too far behind to make any serious run on the Core 2 market except for the fanbois.

This seems like too little too late, to me. But then again I am not in the market for another build. *shrug*

Since I stuffed my Q6600 rig into a 2U case, air cooling was not an option. I typically see the cores running at 25 - 40C with an ambient of 30 inside the case, and (summer) 28C outside the case, this under a full load. (4 core Prime stress) In winter the cores never get above 30C under the same load.

Re:38 C ain't that hot

[Anonymous Coward]

How do you measure the exact amount of current going to the cpu vs the rest of the computer components?

Re:38 C ain't that hot

[Anonymous Coward]

Sorry, that is bullshit. Both Intel and AMD report the same number: maximum power dissipation, since TDP refers to the thermal design power which the cooling solution has to support. "Average" numbers would be useless, since the cooling solution would not be guaranteed to support the peak hotspots which happen even during normal operation.

It would be helpful if some people actually knew what they were talking about...

Re:38 C ain't that hot

[Agripa]

I really don't understand why anyone would buy the Phenom.

AMD has inexpensive ECC support and less expensive motherboards and RAM if you use DDR2. For the difference in price for an equivalent Intel system, I bought a fast hardware RAID controller and 4 drives.

Re:38 C ain't that hot

[toddestan]

If you actually knew what you're talking about, you would know that Intel's numbers are not the worst-case scenarios. It says so right on the processor data sheets. Instead, they "reflect Intel's recommended design point", whatever that is. If you don't believe me, go look for yourself:

http://download.intel.com/design/Pentium4/datashts/29864312.pdf (Random P4 datasheet, look at the footnote at the top of page 70)

Basically Intel relies on their CPU throttling to keep the CPU from overheating under the worst case conditions, or from people installing CPU coolers that exceed their recommendations.

Re:38 C ain't that hot

[Kyont]

Sure, a little slow to respond, but I'm game.

CPU (topic of article): AMD Phenom II X4
CPU cooler: AMD factory fan that came with the CPU
Mobo: ASUS M4A78T-E
Memory: CORSAIR XMS3 DHX 4GB (2 x 2GB) DDR3 memory (motherboard accommodates 4 X 4 GB)
Hard Disk: SAMSUNG Spinpoint F1 HD103UJ 1TB 7200 RPM, SATA connector
Case: Rosewill R5605-BK (two 120 mm fans in front and back)
Power Supply: Rosewill Green Series RG530-2 530W
No additional graphics card or sound card (one of these days though; easy upgrades if cash allows)

Extras:
Card reader/USB: Rosewill RCR-IM5001 USB2.0 75 in 1 internal Card Reader
DVD Burners (x2): SAMSUNG Black OEM (not sure of the model number) with IDE connectors

Not even close to being stuffed full of goodies, but plenty for my moderate needs. I haven't tried overclocking anything. What I've learned from this thread is that I really should add a Kill-A-Watt or other meter to my toolbox. People have a good point about any watts drawn eventually turning into heat. The motherboard is running a little warmer, too (about 45 C) which is a little worrisome and I'll have to keep an eye on. I haven't looked into exactly where on the mobo that measurement takes place.

I have AMD's "Cool 'n' Quiet" setting on, which I think is just marketing speak for "tries hard not to crank when nothing is happening", and all fans (1 CPU and 2 case) on the "Optimal" setting which lets the mobo control them according to temperature (the alternatives are "Performance" i.e. run fans all out all the time, and "Quiet" which means cook your chips as much as possible while trying not to run the fans). I add this info just for the benefit of newbies (it was new to me when I started building) but I imagine it is old hat to many people here.

Re:38 C ain't that hot

[Kyont]

Excellent point. Good old First Law of Thermodynamics, it never steers us wrong! What I've learned from this discussion is that a watt meter would be an essential addition to my tool box. It would also be nice to know how close to its limits the power supply is. It will be twenty bucks well-spent.

Admittedly, the 38 C was under pretty moderate load, probably "typical use". Bunch of apps open, lots of reads and writes, network activity, music playing, web surfing, etc. But nothing like running SETI@home while tarballing my music collection and running a 3-D flight simulator; I'm sure it would get warmer in there. Also admittedly, the motherboard seems to be running warmer than this, around 45 C. I don't know where this is measured or when it gets to be a real concern, but it's high enough to make me uncomfortable. I currently have the fans (1 CPU fan and 2 case fans) set to let the motherboard decide their speed ("Optimal"), but have the option ("Performance") to run them full-speed all the time.

On the other hand, the thing is currently crammed under a desk with poor airflow, and the fans still don't spin very hard to maintain these moderate temps. So I have some hope that when I get it out in the open, I can keep the CPU under say 40 C, even if I have some load spikes or add in more memory or a graphics card.

Re:38 C ain't that hot

[Kyont]

Hey, my boss kept saying we needed an IT Rock Star! It is getting kind of expensive, though, having to build a new machine after every spectacular, caffeine-fueled shutdown.

that's only $35 less...

[pointbeing]

...than I I paid for my i7 920, which is currently running at 3.5GHz without a hiccup.

Now the *motherboard* was considerably spendier ;-)

The deciding factor in Phenom vs Core2

[Sloppy]

I'm trying to pick components now for at least 3 new boxes, and here's my take: All the processors are bloody fucking fast. They're all way faster than what I need. What really matters to me are drivers for low-powered IGP chipsets.

Intel X4500HD vs AMD/ATI G780. Who will get their damn xvmc drivers working first? That is all that matters, and yet it determines what motherboards I'll end up using, therefore CPUs too.

It'll be a damn shame if I cheese out and go with Nvidia. If I have to buy from them just to get things to work, I'm gonna look long and hard to buy some CPUs from just about anyone other than you two assholes. AMD and Intel, it's inexcusable for you to point me at Nvidia. If I have to go that way, I'm gonna harbor a hateful grudge for a long time.

Finish your drivers!

Re:The deciding factor in Phenom vs Core2

[Anonymous Coward]

"It'll be a damn shame if I cheese out and go with Nvidia. If I have to buy from them just to get things to work, I'm gonna look long and hard to buy some CPUs from just about anyone other than you two assholes. AMD and Intel, it's inexcusable for you to point me at Nvidia. If I have to go that way, I'm gonna harbor a hateful grudge for a long time." Unfortunately, VIA's IGPs are probably worse than Intel's or AMD's. Oh well, just have to wait.

Re:The deciding factor in Phenom vs Core2

[julesh]

It all depends on your workload, really. I'm running two similarly specced quad core machines here, one a Phenom 9550, the other a Q6600, so admittedly we're not talking latest-generation stuff, but the difference in the two is quite startling: the Phenom is, I'd say, about 30% faster at handling large vector floating-point tasks (specifically, neural network training is what I'm running on it). The Q6600 is a similar amount faster at more traditional tasks (specifically, I notice this in using a MySQL server under heavy load).

Great value analysis

[Anonymous Coward]

The Tech Report's review has a few more benchmarks, and their unique scatter graphs illustrate the value angle of over 20 CPUs from both Intel and AMD.

Spot the reference? I can play that game.

[argent]

You are Greg Bear and I claim my budget Omphalos in Green Idaho and artificial Tourette's Syndrome.

Re:Spot the reference? I can play that game.

[For a Free Internet]

Close. I am his evil twin who writes Harlequin romance novels. You win one (1) rusty door hinge.

Phenom II X4 965

[Siggy200]

Just purchased a new motherboard, Gigabyte GA-MA790XT-UD4P with a Phenom II X4 965 3.2 GHz chip yesterday. Windows XP Professional seems to load faster on this new board. In the process of installing all my software again.

Re:Phenom II X4 965

[jdb2]

I've been a Linux user for several years and prior to that I was a Windows junkie. Back then, when I upgraded my hardware I never had to do a reinstall of the OS ( XP ) or applications. I don't know if that's what you intended but if not here's a tip :

Say your system files/data are stored on a single hard drive, ( this also applies to the multiple hard drives ) then, before you do anything with regards to a hardware upgrade, such as changing the mainboard, go into the device manager and replace all your system specific devices with the generic Microsoft equivalents. Now, after you've built the new system install your system drive and you should be able to boot up in a generic VESA display mode in which you can allow Windows to autodetect the new hardware and/or in which you can manually install the new system specific drivers.

Always worked for me.

Cheers,

jdb2

Food for processor.

[Anonymous Coward]

Only problem is when your processor gets really hungry it might decide to eat you.

AMD vs Intel

[sanosuke001]

I know I'll be called a fanboy or something but I've been building AMD/ATI systems for the last ten years based solely on the fact that it isn't Intel/nVidia. I'm not going to pay $1000 for the top-of-the-line Intel chip anyway and I'd rather see AMD in business than Intel be the only big player.

Re:AMD vs Intel

[MikeBabcock]

My latest system is an AMD PhenomII with an NVidia card, so how's that for unbiased? :-)

Re:AMD vs Intel

[citylivin]

A fool is someone who blindly sticks to their brand preference. Yes, ten years ago when the athlon xp line of amd processors came out, they were owning up and down the crappy p4s. Intel has come a long way. They pioneered multiple cores, at which time they surpassed amd in terms of performance and heat. Anything before core2 was crap, but after that everyone who doesn't want to lie to themselves switched to intel.

I ran dual opterons for years, but it got replaced with a quad processor intel machine last year. I even had a cyrex back in the day, so I am no fanboi. The better decision is to use what benchmarks the best at the right price point.

As for video cards, ati has always had ass drivers, whereas nvidia drivers are constantly rock solid. For me that is no contest. Being brand loyal is at best a disservice to yourself, and at worst you will make bad purchasing decisions for others. Perhaps if you supplied reasons for your choices, a reasoned argument why ATI and AMD is better (and there are some, price for instance) then people will take you more seriously.

Re:AMD vs Intel

[Anonymous Coward]

Since when is Intel paired with nVidia? I've been building Intel/ATI systems for years!

Re:AMD vs Intel

[Anonymous Coward]

A fool is someone who blindly sticks to their brand preference. ...

As for video cards, ati has always had ass drivers, whereas nvidia drivers are constantly rock solid.

Unless you're talking Linux, this hasn't been true for some time. I've used both nvidia and ATI cards on Windows in games and in the past 4-5 years, ATI has been more consistently stable for me than nvidia. Add in the much better multi-monitor support and I'll generally favor ATI over nvidia in most cases.

Re:AMD vs Intel

[Anonymous Coward]

. . . . and a fool is someone who doesn't try to ensure that he always has options.

Re:AMD vs Intel

[cbhacking]

Regarding drivers, nVidia's drivers were the only part of my Vista system (RC1 to nearly 6 months after RTM) that were anything but rock-solid stable. They crashed (fortunately Vista could restart them without bringing down the OS, but still) and that was after a number of the features were intentionally disabled. Granted, this was on a laptop, but their desktop drivers were almost as bad.

By comparison, at Vista's beta 2, their drivers were already completely solid and provided all the features I expected. Sure, wayyyy back in the day ATi drivers were crap, but on a modern system I'd much prefer ATi drivers over nVidia. Hell, even today on my Win7 RC system, the nVidia driver *claims* to support PhysX, but it doesn't actually *work* when software tries to use it. I'm afraid nVidia has lost all claim to the "superior drivers" card in my book.

Re:AMD vs Intel

[sanosuke001]

Fair enough.

Intel chips for their top of the line are way too expensive. Their chips that are comparable with AMDs top of the line aren't enough of a boost to make switching to Intel worth it; helping the underdog is worth that small difference to me. I also like how the AM2 chipset has a very decent upgrade path; get an AM2+ board and you can run chips from anywhere in the last 2+ years from AMD.

As for ATI and nVidia; there isn't much of a difference between the two. nVidia usually has a small lead on ATI but ATI usually comes out with the top of the line before nVidia comes out with their rebuttal. I have a Radeon 4870 X2 which was the best card out there until the 295 was released. ATIs cards are usually a bit cheaper as well. I've also never had problems with ATIs drivers. The last nVidia card I bought was a geForce 2 ultra (amazing card for $400. lasted me 2-3 years) Again, the only real deciding factor between the two that actually mattered was brand and ATI is considered the underdog.

I will, however, say that I would have liked to have seen AMD support triple channel DDR3 with the phenom II chips but I can see why they might have waited. At the moment, they don't give enough of a performance boost to warrant the extra R&D. It would have been nice to have the option, but dual channel DDR2 does keep costs down.

Re:AMD vs Intel

[Totenglocke]

What card do you have? I've used Win 7 since the beta and never had an issue with drivers on my dual Geforce 9800 GT's on my desktop or my Geforce 7000M on my laptop.

On the subject of AMD vs. Intel, yes, I know that Intel does have processors that are faster. However, you typically have to pay much more for the Intel. Also, I will always support AMD because of Intel's shady business practices. If Intel hadn't spent so many years paying companies like Dell to only use Intel, AMD would've had a lot more sales and would probably have even better technology (before you diss AMD, they were the first ones with 64-bit desktop systems, and if I remember correctly, the first one with dual-core desktop processors as well).

Am I making the best purchase based on bang for my buck all the time by buying AMD? Probably not. However, I'd rather see my money go to support a reputable company than one who has to resort to tactics worse than Microsoft's to be at the top of the industry.

I am out of this game.

[bucky0]

I'm embarassed to say this, but I'm really out of this game. I want to upgrade my machine, and have no idea of what I should get. I want to get a new mobo/cpu/ram combo to put in my machine. Should I wait till more AM3 motherboards come out, driving the price down a bit? I could spend about $400. I do a lot of development (multiple VMs) and play games like TF2 on occasion. Is going to a Core 2 Quad/i7/Phenom II a better choice?

Thanks!
Bucky

Re:I am out of this game.

[Arimus]

For multiple VMs memory is the main need and DDR3 memory is still a premium (not as high as it was but still slightly above) over DD2 memory so an AMD2+ motherboard and an Phenom II 940X would do a lovely job :)

Re:I am out of this game.

[CajunArson]

If you are doing VMs and games, I'd wait a month for the i5 systems from Intel to hit the market. These are the more-affordable mainstream versions of the i7 that are still faster than the AMD boxes while being lower-priced that even a i920, plus the motherboards will be simpler and less expensive than the X58 boards for current i7's. With 4 cores and good power characteristics, you should be fine. The problem with the newest AMD's, aside from power consumption, is that AMD has nothing new planned until at least the end of 2010 aside from maybe . It is very unclear whether or not Bulldozer is going to work on an AM3 platform, while there is a very clear upgrade path for the 1156 sockets through at least 2011.

Re:I am out of this game.

[Fantastic Lad]

Easy. I don't think the following game plan has changed much in the last ten years. . .

When building a tower PC, simply buy whatever fell off the cutting edge six months ago. It'll be a bargain compared to today's new hotness, (which fits your budget), and if it's been long enough for you to no longer know the game, then whatever you end up getting will make your last computer feel like a broken down turtle.

-FL

Bang for the buck

[madwheel]

I just finished my build using AMD for a media server for my house. I know it's not the fastest, but I'm using an OC'd Phenom 9750, MSI MB ATI Radeon 4850, with 4 16x pci-e slots, 4GB of G-Skill PC8500 RAM, 1TB WD Black Edition HD, Antec P180 case, 700watt TT PSU, etc for about $550!! I'd like to see a comparable Intel system put up similar builds for that price. I agree this isn't their best chip for the money but they will always have my support. Without them Intel would not be what they are today.

Re:FAIL?

[V!NCENT]

Depends... OpenCL is comming. I don't see Intel getting massive computations done in the timeframe a GPU can do it. The only one having the best OpenCL performance and at the same time the lowest price and power comsumption per TeraFLOP is AMD with both decent CPU's and GPU's. AMD is also the only one to have the/a full flatform for it. Intel and nVidia can forget it.

Law of Diminishing Returns

[Khyber]

"Many of us thought that we would never see another 140 watt processor (as the Phenom 9950 was), but unfortunately those days are back."

Guess what? In about 15 years, due to the law of diminishing returns we'll be stuck with 200 watt processors for any intensive task like gaming.

What's the latest and greatest processor on a video card draw? Almost 160 watts for my 9800GTX+.

buzz word BS

[cinnamon colbert]

from the story "AMD further attempts to dethrone the Core 2 Quad as the premier midrange CPU offering. While it may not be a world-beater by any stretch of the imagination, it certainly is catching Intel's attention in the breadbasket of the CPU market. "
?premier midrange cpu offering - talk about buzzwords strung together
and how is this consistent with breadbasket ?

The Core i5 is the real competitor to this thing.

[CajunArson]

Tom's Hardware's review points out that the real competitor to this chip is the Core i5 750 which is coming out next month and from Tom's simulations using a downgraded Core i920 it looks like the AMD chip is already beaten.. meaning that AMD is going to have to cut the price on this thing quickly. Basically: If you want one of these things, wait a month until the i5's become available and then buy them for $200 where AMD will be forced to price them. Or, if you are building a whole new system, I'd get an i5 since it combines most of the performance of the i7 with a lower cost structure and future-upgradeability which AMD cannot offer right now. The Phenom II's had a decent if short run as competitors to the Core 2. Unfortunately for AMD the Core 2 is heading for bargain basement territory by the end of the year, and AMD (which desparately needs a new microarchitecture) has nothing to offer but overclocks of an already hot chip until at least last 2010 if not 2011.

Total system power draw only reliable measurement

[Anonymous Coward]

Measuring the power of an individual component is highly error prone and is inaccurate at best.

Measuring the total system power draw is the only reliable measurement.

If the AMD system under load uses only 15W more than the Intel one with the same or very similar components, but costs much less, that's a reasonable compromise.

OpenCL is coming

[symbolset]

And so is Larrabee. The next year should be interesting in the hobby number crunching game.

Re:OpenCL is coming

[V!NCENT]

I hope for Intel that Larrabee's power can be fully utilised by OpenCL, because what makes Larrabee so Larrabee is it's specialised instruction set. And who is going to want to code solely for Larrabee if they can code for everything with OpenCL?

I won't buy Intel on moral grounds

[Totenglocke]

AMD has never had to resort to paying other companies off to stay ahead of the competition. Intel however has used this tactic for years. As a result, I'll never buy Intel because I refuse to pay money to such a company like that. I'd rather my money go to AMD, who is honest, and every time I buy AMD it gives them more money to develop newer, better processors.

I hate to say it...

[CAIMLAS]

I hate to say it, but as someone who just bought a Phenom II x3 710 and someone who has exclusively bought AMD processors (for non-laptop use) for personal use since the K6, AMD may have just lost a customer for future purchases.

AMD: you need to stop focusing on the 'performance' side of things: people who look at performance tend to look at benchmarks and relative value. You need to put heavy focus on making your dies smaller and power envelope lower, while trying to regain memory support and bus superiority again. That was what I looked for when I bought, and those are the things that the common consumer look for (albeit usually in terms of "power use", "cost", and "snappy response").

Motherboard swap

[Anonymous Coward]

I have to agree. My method is to install the drivers for the new motherboard and then swap them out. With this method, I've never spent more than 15 minutes configuring the new drivers. I've swapped motherboards probably 10 times using this method (without any BSOD's)