Phenom VS. Intel Quadcores (Q6600) ?

NathanP2007

[H]ard|Gawd
2FA
Joined
Dec 17, 2007
Messages
1,697
OK you guys dont need to go into major detail, but im just curious for when i upgrade to quad core (got a AMD X2 5200+) which one is better for gaming? Iv heard that AMD is better for gaming, but i have also heard that Intel is better for newer technology and better Quad cores and better for whats new on the market. What do you guys think?
 
if you already have AM2, it would be better to buy a Phenom over a Q6600 unless you want to buy a new motherboard...
 
getting a new mobo isnt a issue, im just curious about which is better for gaming / better in general
 
I went from an X2 5000+ to a q6600, I love this thing, works great, overclocks way better than any AMD I've owned (not the phenoms though).
 
Not to mention that most Q6600 Revision G0's are getting 40-60% overclocks on air.
Phenom: not so much...
 
The Q6600 is better, there is no disputing that fact, some guys might come in here talking about the next revision, or a lower price etc etc, but the Q6600 is faster than ANY AMD processor.
 
If you are not into OC or just wanna OC a bit, you can't go wrong with either CPU but if you wanna OC like mad, the Q6600 with definitely reach a higher speed.

For most of the time you won't notice the performance difference actually because the Phenom is fast enough even at stock, and for games, the extra performance of the CPU will still be limited by the GPU anyway. When it is not limited by the GPU, you will most likely reach 100+ fps already so I doubt that you can tell the difference between 120fps and 150fps.

Unless you are doing some serious video encoding and such, you won't use all the processing power much actually. What I like about the Phenom is that with Cool n Quiet enabled, each of the core can have different clock speed depending on the load. Since most of current softwares are still single threaded so in this situation, only one core will be in full speed and the other three will be in an energy saving mode.

What you must remember is that when you will pay a cheaper price for the Phenom so in return you will get less performance compared to the Q6600.
 
if you already have AM2, it would be better to buy a Phenom over a Q6600 unless you want to buy a new motherboard...

Only if your AM2 board actually supports Phenom, which a lot don't!
If you're going to buy a new motherboard anyway, you might aswell go Q6600, especially if you're overclocking (and if not, what are you doing on this forum? :))
 
For most of the time you won't notice the performance difference actually because the Phenom is fast enough even at stock, and for games, the extra performance of the CPU will still be limited by the GPU anyway. When it is not limited by the GPU, you will most likely reach 100+ fps already so I doubt that you can tell the difference between 120fps and 150fps.

I think that's nonsense, really.
"It's fast enough". Don't you wan't your money's worth? Don't you want the best performance possible, so you will have it by the time you DO notice? It may allow you to postpone your next upgrade by 6 months to a year.

What I like about the Phenom is that with Cool n Quiet enabled, each of the core can have different clock speed depending on the load. Since most of current softwares are still single threaded so in this situation, only one core will be in full speed and the other three will be in an energy saving mode.

It doesn't seem to do AMD any favours in practice. Power consumption of the Phenom is still about the same as a Q6600, even at stock: http://www.hardocp.com/article.html?art=MTQyMiw5LCxoZW50aHVzaWFzdA==
Considering that the Q6600 is quite a bit faster, especially with the TLB fix installed (these benches were without the fix), Q6600 wins in terms of performance-per-watt. If you are really bothered about power consumption, you'd have a more efficient system with a Q6600 that is slightly undervolted and/or underclocked (the overclocking potential of Q6600 goes both ways... because it can run much faster than stock even at stock voltage, you can run it at stock speeds at a much lower voltage aswell).
But as you can see, the real performance-per-watt king is going to be the QX9650.
 
I think that's nonsense, really.
"It's fast enough". Don't you wan't your money's worth? Don't you want the best performance possible, so you will have it by the time you DO notice? It may allow you to postpone your next upgrade by 6 months to a year.

It doesn't seem to do AMD any favours in practice. Power consumption of the Phenom is still about the same as a Q6600, even at stock: http://www.hardocp.com/article.html?art=MTQyMiw5LCxoZW50aHVzaWFzdA==
Considering that the Q6600 is quite a bit faster, especially with the TLB fix installed (these benches were without the fix), Q6600 wins in terms of performance-per-watt. If you are really bothered about power consumption, you'd have a more efficient system with a Q6600 that is slightly undervolted and/or underclocked (the overclocking potential of Q6600 goes both ways... because it can run much faster than stock even at stock voltage, you can run it at stock speeds at a much lower voltage aswell).
But as you can see, the real performance-per-watt king is going to be the QX9650.
You want the best performance, buy the QX CPU and see your $1000 be wasted. Like I said before the Phenom is cheaper so it will have less performance, but most of the time you won't notice the difference, even in games. I don't know about you but for me, I would rathe use a higher resolution and image quality in games even if my fps will be only at 60fps than using a lower resolution with less image quality to get 100+ fps. With higher resolution and higher image quality, you will be limited by your GPU anyway.

What most benchmarks show is its power consumption in full idle and full load(all cores), not the power consumtion during only one core utilisation which is most of the current softwares will do. We have ran two Phenoms so far and never encountered any instability due to the TLB bug because the bug will only affect the virtual machine usage which most the user will not use anyway. Unless you have used the Phenom yourself, please stop spreading BS because I'm getting sick of it.

Let see again the numbers:
Q6600 100% Load: 160W
Q6600 75% Load: 156W ======9600 100% Load: 155W
Q6600 50% Load: 150W ======9600 75% Load: 151W
Q6600 25% Load: 137W ======9600 50% Load: 140W
========================9600 25% Load: 128W
Q6600 0% Load: 108W =======9600 0% Load: 105W

I wonder why X-Bit labs left out the 9600 and 9500 in their power consumption test :rolleyes:
 
You want the best performance, buy the QX CPU and see your $1000 be wasted. Like I said before the Phenom is cheaper so it will have less performance, but most of the time you won't notice the difference, even in games. I don't know about you but for me, I would rathe use a higher resolution and image quality in games even if my fps will be only at 60fps than using a lower resolution with less image quality to get 100+ fps. With higher resolution and higher image quality, you will be limited by your GPU anyway.

What most benchmarks show is its power consumption in full idle and full load(all cores), not the power consumtion during only one core utilisation which is most of the current softwares will do. We have ran two Phenoms so far and never encountered any instability due to the TLB bug because the bug will only affect the virtual machine usage which most the user will not use anyway. Unless you have used the Phenom yourself, please stop spreading BS because I'm getting sick of it.

Read my link please, it shows AMD preformance is sub-par...unlke the rigged AMD-PR-Benches
 
Read my link please, it shows AMD preformance is sub-par...unlke the rigged AMD-PR-Benches

I wonder why X-Bit labs left out the 9600 and 9500 in their power consumption test :rolleyes: So other benches are rigged by AMD because the benches show that the difference is not that big but X-Bit labs is not biased when they left out the 9600 and 9500 in their power consumption test?
 
This will answer the question:
http://www.xbitlabs.com/articles/cpu/display/amd-phenom.html

No guessing, just the facts and not AMD's rigged benchmarks...from Phenom 9500 to a Phenom @ 3.0 Ghz...it dosn't look pretty for AMD...

Who plays games at 1024x768?

Besides [H}ardOCP's review shows the 9700 trading blows with the Q6600 in gaming.

And if they are with in ~5fps at low resolutions the difference at a high res is even going to be less noticable.

Sure the Intel chips overclock easier and higher, not arguing that point, but there are alot of people that dont overclock.

But I guess we will have to wait for the [H]'s next real world gaming comparison.
 
I wonder why X-Bit labs left out the 9600 and 9500 in their power consumption test :rolleyes: So other benches are rigged by AMD because the benches show that the difference is not that big but X-Bit labs is not biased when they left out the 9600 and 9500 in their power consumption test?

They where rigged due the the HT-bus difference ;)
 
Who plays games at 1024x768?

Besides [H}ardOCP's review shows the 9700 trading blows with the Q6600 in gaming.

And if they are with in ~5fps at low resolutions the difference at a high res is even going to be less noticable.

Sure the Intel chips overclock easier and higher, not arguing that point, but there are alot of people that dont overclock.

But I guess we will have to wait for the [H]'s next real world gaming comparison.

I think a lot of people still play games at 1024x768 so the real question is who buys a quad core CPU but stll plays games at 1024x768 :p
 
Who plays games at 1024x768?

Besides [H}ardOCP's review shows the 9700 trading blows with the Q6600 in gaming.

And if they are with in ~5fps at low resolutions the difference at a high res is even going to be less noticable.

Sure the Intel chips overclock easier and higher, not arguing that point, but there are alot of people that dont overclock.

But I guess we will have to wait for the [H]'s next real world gaming comparison.

1024x768 is to make it CPU limited, going higher would make it GPU limited, please educate yourself.
 
FWIW I'm a long time AMD fanboy, and I chose to make the move to penryn in the next month or so with the release of the Q9450 over going for Phenom.
 
1024x768 is to make it CPU limited, going higher would make it GPU limited, please educate yourself.

No shit you dont say, wow didnt know that.

Other then maybe Crysis and FSX dont we all play our games with the settings to the point in which we are GPU limited. Well I do and if you dont maybe you should go edjucate yourself. :rolleyes:
 
You want the best performance, buy the QX CPU and see your $1000 be wasted. Like I said before the Phenom is cheaper so it will have less performance, but most of the time you won't notice the difference, even in games. I don't know about you but for me, I would rathe use a higher resolution and image quality in games even if my fps will be only at 60fps than using a lower resolution with less image quality to get 100+ fps. With higher resolution and higher image quality, you will be limited by your GPU anyway.

Well, I tend to upgrade my GPU more often than my CPU. If I were to do that on a Phenom I'd be CPU-limited sooner than on a Core2 Quad system. So I repeat: get the best performance for your money, it may allow you to get another 6 to 12 months out of your system. Better long-term investment.
I bet you weren't telling the same nonsense back when Pentium 4 was slower than Athlon64?
Pentium 4 was still 'fast enough' for most applications anyway.
Useless fanboy drivel.

What most benchmarks show is its power consumption in full idle and full load(all cores), not the power consumtion during only one core utilisation which is most of the current softwares will do. We have ran two Phenoms so far and never encountered any instability due to the TLB bug because the bug will only affect the virtual machine usage which most the user will not use anyway. Unless you have used the Phenom yourself, please stop spreading BS because I'm getting sick of it.

Let see again the numbers:
Q6600 100% Load: 160W
Q6600 75% Load: 156W ======9600 100% Load: 155W
Q6600 50% Load: 150W ======9600 75% Load: 151W
Q6600 25% Load: 137W ======9600 50% Load: 140W
========================9600 25% Load: 128W
Q6600 0% Load: 108W =======9600 0% Load: 105W

Yes, that was what I was referring to. As you can see the difference about 7W on average, across the board.
At a total of 160W, that is about 4% more power usage. However, in the benchmarks, the Q6600 is more than 4% faster than the 9600 (and even the 9700) on average, hence, the Q6600 has the best average performance-per-watt. As I already said.

Also, just drop the voltage a bit on the Q6600, and you'll easily wipe away the 7W more power consumption, without sacrificing the performance.
So again I'm just repeating what I already said, because you aren't actually making any kind of point, really.

Also, I'm not spreading BS, it's AMD who spread the information about the TLB bug. I'm just saying that if you have the fix installed (which most people probably will, with a motherboard with the latest bios), you will get even lower performance than in the published benchmarks.
I never said you had to enable the fix, I never used the word 'instability' at all. These are just the facts.
However it IS BS that it only affects virtual machines. This is simply not true. Virtual machines are just more prone to the bug, because they use the TLB more extensively. Stop spreading lies.
 
What HT-Bus difference?

The pre-production chips sent to reviewers (such as the one used in the Hardocp and X-Bit reviews) had a 2 GHz HT-bus clock. Production models have 1.8 GHz HT-bus clock.

That's why I said that most of the time you will not notice any difference even in games.

Why 'even'?
Games are the exception to the rule. Most other applications don't make use of a GPU, so they rely on the CPU only. Which is why you notice it A LOT in 3d rendering, audio/video encoding etc.
You could argue that you don't notice it in Office or browsing... but then why are we arguing about quadcores in the first place? Might aswell take a dualcore for that, they can't be beat in terms of price and performance per watt.
And as I already stated... upgrade to the next generation of GPUs, and you'll probably find yourself CPU-limited again.
 
The pre-production chips sent to reviewers (such as the one used in the Hardocp and X-Bit reviews) had a 2 GHz HT-bus clock. Production models have 1.8 GHz HT-bus clock.



Why 'even'?
Games are the exception to the rule. Most other applications don't make use of a GPU, so they rely on the CPU only. Which is why you notice it A LOT in 3d rendering, audio/video encoding etc.
You could argue that you don't notice it in Office or browsing... but then why are we arguing about quadcores in the first place? Might aswell take a dualcore for that, they can't be beat in terms of price and performance per watt.
And as I already stated... upgrade to the next generation of GPUs, and you'll probably find yourself CPU-limited again.

So you think that most user will do a lot of 3D rendering and multimedia encoding? Most people buy the quad core because it is cheap enough to buy it, only a small percentage of the quad core user really do 3D rendering and encoding, most of the time they think that with a Q6600, they will get a better performance in games. The current gen GPUs are hardly limited by AMD Athlon X2, let alone C2D, by the time you will be limited by the Phenom, you will most probably looking for a new system with DDR3, DirectX 11 and PCI-E 3.0 already.
 
So you think that most user will do a lot of 3D rendering and multimedia encoding? Most people buy the quad core because it is cheap enough to buy it, only a small percentage of the quad core user really do 3D rendering and encoding, most of the time they think that with a Q6600, they will get a better performance in games.

I never claimed a Q6600 will get better performance in games... at least, not with current GPUs.
I also never recommended anyone to buy a quadcore.
So I don't see what this has to do with me.

All I say is that if you're going to buy a quadcore anyway, you might aswell get the best performer in that pricerange and get your money's worth.

The current gen GPUs are hardly limited by AMD Athlon X2, let alone C2D, by the time you will be limited by the Phenom, you will most probably looking for a new system with DDR3, DirectX 11 and PCI-E 3.0 already.

Don't know if it's because I've been in this business much longer than you, or because developing graphics software is what I do... but my experience tells me completely different.
People thought the same only a year ago... just before the GeForce 8800 series hit. *boom*, all of a sudden a fast CPU was really important for gaming. And I can mention numerours other examples of such. A generation of graphics hardware tends to last only 1.5 to 2 years tops... And the 8800 is already a year old. nVidia is doing a sort of tick/tock aswell. They've just shrunk their 8800... next up is a new architecture... Probably not yet the upcoming 9-series (although... triple SLI?), but its successor, which will hit us in 6-12 months. By then we'll have at least twice the GPU power we have today.
 
Don't know if it's because I've been in this business much longer than you, or because developing graphics software is what I do... but my experience tells me completely different.
People thought the same only a year ago... just before the GeForce 8800 series hit. *boom*, all of a sudden a fast CPU was really important for gaming. And I can mention numerours other examples of such. A generation of graphics hardware tends to last only 1.5 to 2 years tops... And the 8800 is already a year old. nVidia is doing a sort of tick/tock aswell. They've just shrunk their 8800... next up is a new architecture... Probably not yet the upcoming 9-series (although... triple SLI?), but its successor, which will hit us in 6-12 months. By then we'll have at least twice the GPU power we have today.

http://www.firingsquad.com/hardware/geforce_8800_gtx_gts_amd_cpu_scaling/

So much of a clock work:
http://www.hardforum.com/showthread.php?t=1259516
Just like Intel's clock work, messed up the P35 45nm support and put the blame on AMD.
 
Is there anything you'd like to say with these links?
Because they seem to say exactly what I've already said in my post above.

You see that when you are not limited by the GPU in older games, you will get a 100+ fps already so why don't just use a higher image quality and resolution? Newer games will be limited by the GPU first, not by the CPU. I don't know, but I don't think that getting 150fps instead of 120fps makes a difference. When you are only getting 60+fps, the CPU doesn't make any difference at all.
 
You see that when you are not limited by the GPU in older games, you will get a 100+ fps already so why don't just use a higher image quality and resolution? Newer games will be limited by the GPU first, not by the CPU. I don't know, but I don't think that getting 150fps instead of 120fps makes a difference. When you are only getting 60+fps, the CPU doesn't make any difference at all.

I was talking about current and future games (think Crysis), and future GPUs (the next big thing after the 8800 series). In a year or so we'll have more of these games, and newer GPUs which can reach much higher framerates (Crysis doesn't get anywhere near 60+ fps on any system), and require much faster CPUs, because also physics, AI and other stuff has to be processed at higher framerates now.
As the benches already show on these low resolutions... the Q6600 does indeed make quite a difference compared to a 9600 in games, if the GPU is fast enough.
So when we get cards that can run Crysis at 60+ fps, you'll be needing CPUs that can run Crysis at 60+ fps aswell. Q6600 is the one, just look at the difference, it's over 20% faster: http://www.xbitlabs.com/articles/cpu/display/amd-phenom_11.html#sect0
Not even the Phenom overclocked to 3 GHz can manage anywhere near 60 fps... a Q6600 at 3 GHz will.

Everyone who buys a Phenom 9600 now, and gets a new high-end GPU in 6 months is going to really slam their faces into the wall for not going Q6600 when they're running Crysis (or any other next-gen game that may be out by then).
The difference is quite large, and gets even larger when you start overclocking.
 
What HT-Bus difference?

The 2.0Ghz rigged benchmarks vs the real world 1.8Ghz bus speed...if Intel had done this, you be crying like a virgin for he first time, but are you now saying you didn't know of this..and yet you give advice...without the proper knowlegde?


That's why I said that most of the time you will not notice any difference even in games.

Like Scali2 has said, the faster the GPU the more the CPU will matter...but if you to save a miniscule buck can live with subpar preformnace, more power usage, bad OC ect...by all means..go buy a Phenom.
 
The 2.0Ghz rigged benchmarks vs the real world 1.8Ghz bus speed...if Intel had done this, you be crying like a virgin for he first time, but are you now saying you didn't know of this..and yet you give advice...without the proper knowlegde?




Like Scali2 has said, the faster the GPU the more the CPU will matter...but if you to save a miniscule buck can live with subpar preformnace, more power usage, bad OC ect...by all means..go buy a Phenom.

So you think that the the HT speed would make much of a difference? You didn't answer my question on why would X-Bit Labs left out the power consumption of the 9600 and 9500?
Not to get involved in what seems to be a perpetual spat between you and alg7_munif, but I do want to be realistic here: those benches are done at 10x7, with medium quality.

I guess that he has tried to play Crysis on both CPUs using the same HD3870 like I did.
 
Not to get involved in what seems to be a perpetual spat between you and alg7_munif, but I do want to be realistic here: those benches are done at 10x7, with medium quality.

I know, that is exactly the point.
These benchmarks are similar to benching with a faster future GPU at higher resolutions and higher settings. Obviously we can't actually bench with these GPUs yet, since they don't exist. We just know that they will be faster than current GPUs.
In fact, if they set ALL settings to medium quality, including physics, then high or very high settings will require even more CPU power to reach 50-60 fps. Crysis is just a very CPU-heavy game, if you take the GPU out of the equation for a moment. It won't run 100+ fps on any current CPU, that's for sure. That was the point I was making.
 
So you think that the the HT speed would make much of a difference? You didn't answer my question on why would X-Bit Labs left out the power consumption of the 9600 and 9500?

Ask them, not me :rolleyes:
Let me put this in perspective for you:
Why did AMD make rigged benchmarks (so rigged that HardOCP didn't want no part of it) with 2.0Ghz parts, when they knew that production models only would be 1.8Ghz?

Should I ask YOU this..or AMD? :rolleyes:
 
I guess that he has tried to play Crysis on both CPUs using the same HD3870 like I did.

You don't seem to get it. Obviously with a slow card like that, you'll be GPU-limited.
That doesn't mean that Crysis isn't heavy on the CPU. Just turn down the settings to a minimum, and you'll see that the system still cannot reach 100+ framerates.
Geez, pisses me off... is GPU-limited/CPU-limited such a tough concept to grasp for you non-tech people?
Being limited by one or the other says nothing about the actual framerates you'll be getting.
 
A quote from the man himself

Yea, he talks about *now*, I'm talking about the near future.
As I say, you're likely to upgrade your GPU at least once before you replace your CPU, motherboard and other stuff.
I'm sure if we ask Kyle, he'll say the same thing: as GPUs get faster, CPU speed becomes more important.
He'll probably also agree that you'll NEVER be able to play Crysis at 60+ fps on a Phenom, not even at 3 GHz. The CPU simply isn't fast enough, no matter what GPU or what settings you use.
 
Geez, pisses me off... is GPU-limited/CPU-limited such a tough concept to grasp for you non-tech people?

Guy, chill out. Nobody appreciates being insulted.

I understand the concept behind GPU/CPU limited performance. I think both you and alg7_munif are arguing about the same things; right now there is no appreciable difference (at real-world settings), but once GPU performance increases we will see a return of CPU limited performance.
 
I understand the concept behind GPU/CPU limited performance. I think both you and alg7_munif are arguing about the same things; right now there is no appreciable difference (at real-world settings), but once GPU performance increases we will see a return of CPU limited performance.

I don't think alg7_munif realizes that second part. Else, why would he recommend a CPU that is bound to become considerably slower in games as soon as you upgrade your GPU?
The few dollars price difference really don't account for the large difference in performance that is just around the corner.
Therefore his whole "it doesn't matter in games, so get the Phenom" is rather ill-advised.
 
The few dollars price difference really don't account for the large difference in performance that is just around the corner.

I don't disagree; last I checked, it was a mere $20.
 
I don't disagree; last I checked, it was a mere $20.

Yea, if anything, Intel overestimated AMD this time.
They probably thought that a quadcore at 'only' 2.4 GHz would be a good entrylevel point.
I bet they wish they'd made a 2.13 GHz quadcore aswell, so they could actually compete at the same performance level as AMD, and just beat them with the price.
Then we wouldn't have to hear that "But Phenom is cheaper" drivel all the time. Some people may actually buy into that nonsense.
 
Back
Top