Would you be happier if Kyle used a Q9550 and overclocked it to 3.2GHz? The results would have been the same, except using a $300 CPU instead of a $1500 chip. The point was to have a clock-for-clock comparison, and presumably it was convenient to use the QX9770 since it already ran at 3.2GHz. Any other Yorkfield running at 3.2GHz would have performed identically.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
AMD Phenom II X4 Model 940 @ [H]
- Thread starter FrgMstr
- Start date
Would you be happier if Kyle used a Q9550 and overclocked it to 3.2GHz? The results would have been the same, except using a $300 CPU instead of a $1500 chip. The point was to have a clock-for-clock comparison, and presumably it was convenient to use the QX9770 since it already ran at 3.2GHz. Any other Yorkfield running at 3.2GHz would have performed identically.
- Joined
- May 18, 1997
- Messages
- 55,634
I think it is ludicrous to think you need an expensive Intel chip to do 3.2GHz. Jesus, any of them will do that starting at under $200.
And if you get that impression, you obviously did not RTFA thoroughly.
- Joined
- May 18, 1997
- Messages
- 55,634
If you have a board with good power on it that will handle a solid 140w load or so and OC it to 3.4 or so, I think it would be worth the investment as we spelled out in the article. I would wait for the price to come down though, you know it will.
jonneymendoza
Supreme [H]ardness
- Joined
- Sep 11, 2004
- Messages
- 6,395
sheesh? man when will amd fanboys realize that amd cpu's are no longer any good anymore? the core 2 has been out for two years and its just now amd are just about competing with it.
Tech-Daddy
Gawd
- Joined
- Jul 28, 2004
- Messages
- 514
Disappointing to say the least.
The value prop on an AMD vs Intel is subjective to the end user. I am not a blind "fanboy" for AMD, but chose their CPU for personal reasons.
I'm not a bench mark maniac, I dont need bragging rights.... I need a rig that will let me enjoy the games I want to play, at a reasonable price, and gives me a good bit of forward upgrade path. That is something that has *always* pissed me off about Intel offerings. Different sockets for new CPU's forcing new motherboards... grrrr.... good for their bottom line, but a hard core kick in the nuts to enthusiasts.
Will I have a machine that will win a benchmark duel. Nope... not unless AMD pulls off another Athlon64 stunner, which I think we can all agree is not likely to happen with the current conservative direction that AMD has begun taking.
I I know that I will have a hardware platform that I can keep for several years, moving it down the food chain through the kids computing needs... and have many CPU upgrade options available to me at the drop of a BIOS flash. I cannot say that with any confidence on an Intel platform.
So, "yeah", if you setup an AMD rig, you are not going to be getting the extreme performance of the Intel offerings, be it C2D, C2Q or i7... however the $250 to nearly $400 entry point for a new i7 motherboard, tripchan memory, and the CPU.... gads! No effing way! Not in this economy. Buy C2Q/mobo/memory, and you have just limited yourself on an upgrade path on the Intel side.
*shrug*
It's all perspective and how you spend your money on technology.
I will say this Kyle, you sounded quite pissed off in the article and went out of your way to make sure everyone knew you were not impressed. Aggressive writing, "tell it like it is" stype that [H] has been known for, while most times I have chuckled about (even when bitching out AMD), this seemed more personal. Not sure where it came from, but this seemed more of an attempt to attack AMD's PR machine and kick their respective teeth in, rather than a technical shakedown of their product.
I wish AMD would be competitive again in the high performance arena... with each new release they make... my fears get deeper.
Pretty sooon, if things dont change, I'm going to be gaming on the relative performance equivalent of a Cyrix!
Always informative, always [H]ard.... I stay an avid reader.
-=TD
The value prop on an AMD vs Intel is subjective to the end user. I am not a blind "fanboy" for AMD, but chose their CPU for personal reasons.
I'm not a bench mark maniac, I dont need bragging rights.... I need a rig that will let me enjoy the games I want to play, at a reasonable price, and gives me a good bit of forward upgrade path. That is something that has *always* pissed me off about Intel offerings. Different sockets for new CPU's forcing new motherboards... grrrr.... good for their bottom line, but a hard core kick in the nuts to enthusiasts.
Will I have a machine that will win a benchmark duel. Nope... not unless AMD pulls off another Athlon64 stunner, which I think we can all agree is not likely to happen with the current conservative direction that AMD has begun taking.
I I know that I will have a hardware platform that I can keep for several years, moving it down the food chain through the kids computing needs... and have many CPU upgrade options available to me at the drop of a BIOS flash. I cannot say that with any confidence on an Intel platform.
So, "yeah", if you setup an AMD rig, you are not going to be getting the extreme performance of the Intel offerings, be it C2D, C2Q or i7... however the $250 to nearly $400 entry point for a new i7 motherboard, tripchan memory, and the CPU.... gads! No effing way! Not in this economy. Buy C2Q/mobo/memory, and you have just limited yourself on an upgrade path on the Intel side.
*shrug*
It's all perspective and how you spend your money on technology.
I will say this Kyle, you sounded quite pissed off in the article and went out of your way to make sure everyone knew you were not impressed. Aggressive writing, "tell it like it is" stype that [H] has been known for, while most times I have chuckled about (even when bitching out AMD), this seemed more personal. Not sure where it came from, but this seemed more of an attempt to attack AMD's PR machine and kick their respective teeth in, rather than a technical shakedown of their product.
I wish AMD would be competitive again in the high performance arena... with each new release they make... my fears get deeper.
Pretty sooon, if things dont change, I'm going to be gaming on the relative performance equivalent of a Cyrix!
Always informative, always [H]ard.... I stay an avid reader.
-=TD
There is nothing wrong with what you've got. It is just slower than Intel's offerings in the same price range. As we've said time and time again in gaming applications this won't be readily apparent. Your video card is still the most important piece of hardware in determining gaming performance.
If you feel the performance lacking then of course, a Phenom II might be just the thing for you. If you find yourself wanting a new build, don't let anyone stop you. I'd go with a Gigabyte EX58-UD5 and a Core i7 920. Not a budget build by any means but not a super expensive build either. Overclock that sucker and you'll be good to go for quite some time.
i bet its very competetive clock for clock vs core 2, but this quad extreme have more cache though.
using an 8 mb cache cpu would have altered the score Slightly.
can you clock a Q9400 q9550 to 3.8 ghz on air?
i surely havnt seen it, and you can suicide @ 4.8 ghz on air with good chip.
if you want to have a chip to play with, this looks like a promosing chip, LN2 Lhe Dry ice
phase change.
no cold bug.
gonna be hell loads of fun ;D thats why im getting in, if it hadnt been that overclock anything i see, i wouldnt have bought any new chips at all before i had too, and a x2 5600+ actually works quite nice in games @ 1680x1050 gaming, and havnt shown any signs of lagging more than my phenom @ 3400 mhz.
using an 8 mb cache cpu would have altered the score Slightly.
can you clock a Q9400 q9550 to 3.8 ghz on air?
i surely havnt seen it, and you can suicide @ 4.8 ghz on air with good chip.
if you want to have a chip to play with, this looks like a promosing chip, LN2 Lhe Dry ice
phase change.
no cold bug.
gonna be hell loads of fun ;D thats why im getting in, if it hadnt been that overclock anything i see, i wouldnt have bought any new chips at all before i had too, and a x2 5600+ actually works quite nice in games @ 1680x1050 gaming, and havnt shown any signs of lagging more than my phenom @ 3400 mhz.
jonneymendoza
Supreme [H]ardness
- Joined
- Sep 11, 2004
- Messages
- 6,395
people moan about the price of a i7 when if u do build one, you will get your moneys worth for a long time as nothing will beat it for at least a year if not more
- Joined
- Oct 11, 2005
- Messages
- 4,240
I would buy a phenom 2 for 100 bucks and not a dollar more after reading this. I was hoping for a part that I could simply drop into my motherboard and be good for a while...but it seems after this article that it's worth my while to wait until I have the money to invest in Intel.
Thanks Kyle for this great review. It's perfect and it's eye opening, and people who bitch about it are people who dont like the information that they see.
Thanks Kyle for this great review. It's perfect and it's eye opening, and people who bitch about it are people who dont like the information that they see.
cannondale06
[H]F Junkie
- Joined
- Nov 27, 2007
- Messages
- 16,180
hmm most of the other game benchmarks i am seeing show the Phenom 2 being dead even and sometimes even beating similar clocked Core 2 cpus at real world settings.
- Joined
- May 18, 1997
- Messages
- 55,634
How exactly does the amount RAM impact multiple benchmark runs when the system never pages anywhere near into the amount of RAM on the system?
Please provide stark links of contrast. Please.
Intel's current mainstream Core 2 Quad CPUs all have either 6MB or 12MB of cache. The 8MB models are based on the old 65nm Kentsfield core, which is only used in the Q6600, which, as far as Intel quads go, is a low-end chip.
Yes, you can clock a Q9400 or Q9550 to 3.8GHz on air. My main rig has a Q9550 running at 3.8GHz, and I keep it at full load 24/7 running Folding@Home.
vista_blista
Limp Gawd
- Joined
- Jul 21, 2007
- Messages
- 480
Probably operator error combined with an unrealistic viewpoint. Lap dogs seem to love it so far though.
- Joined
- May 18, 1997
- Messages
- 55,634
When I say that I mean in the smallest sense, not to where it would make any difference really. Minuscule. It has been my experience though. Take system x with 2GB of ram, then System X with 4GB of ram, same exact settings. 4GB numbers are always a tad slower. But in minuscule amounts. Basically I think it is a non-issue.
- Joined
- May 18, 1997
- Messages
- 55,634
The big key there is that you are talking about games that are likely loading huge amounts of data the require more than 2GB to keep from having to go back to the disk. The problem with that article is that it fails to show you WHY the RAM makes a difference in those situations.
Hey Kyle, crew thanks for the review.
I think the problem with [H] numbers or actually with the numbers of the other sites is that [H] is the only sites so far that uses superfast Intel SSDs. On how much that might impact min and avg frame rate I am not sure about but from anand tests faster HDs do improve min fps in games like crysis. And with extremely fast cpus like 45nm core2s @ 3.8 or core i7 @ 3.8, Intel SSD might allow them to generate even higher fps then ever possible with usuall 7200 rpm shittys. Just a hypothesis for you guys to test out if you ever get the time.
I think the problem with [H] numbers or actually with the numbers of the other sites is that [H] is the only sites so far that uses superfast Intel SSDs. On how much that might impact min and avg frame rate I am not sure about but from anand tests faster HDs do improve min fps in games like crysis. And with extremely fast cpus like 45nm core2s @ 3.8 or core i7 @ 3.8, Intel SSD might allow them to generate even higher fps then ever possible with usuall 7200 rpm shittys. Just a hypothesis for you guys to test out if you ever get the time.
Are you sure? Unless someone does compare a same system but with different HD types its still up for debate.
jonneymendoza
Supreme [H]ardness
- Joined
- Sep 11, 2004
- Messages
- 6,395
wow this thread is turning into a love hate one! seems like some people dont like the review here
What is a SSD/HDD going to do in aiding computation of a game? Nothing, it is going to slow it down sure but only when it is being accessed. When you're specifically trying to test a component you work to eliminate such scenarios.
Hypothetically, yes but since [H] does real world gaming tests loading different stuff in the level might slow down core 2s and i7s more then you can imagine when you're using an 7200rpm HD.
Oh, btw Kyle is the 8Gb Intel SSD a typo right? I'm not aware of any 8Gb Intels so Im guessing you guys got the 80Gb version. Nice. Very nice.
- Joined
- May 18, 1997
- Messages
- 55,634
Good one, first time I have heard that one. You get kudos for originality.
Just ran the FarCry 2 benchmark with the largest memory footprint that we used, and it pulled a whopping total system RAM of 1.2GB. Ran it a couple times, same result. The Intel SSD we are using to reduce IO bottlenecks was nice and quiet too after the benchmark full loaded the level.
Hope that satisfies your interest. If you do not understand that, I am not sure how to explain it to you. It just comes down to you being concerned about volume and we are worried more about making sure we had the bandwidth while never hitting close to maxing volumn. Dunno if that makes sense to you or not....???
So I guess, in general, if both are overclocked at the max possible speed, the Q9400 is probably better than the X4 940. Pretty much all the other reviews run the chips at stock, where the 940 is about the same or slightly better than the Q9400 and worse than the Q9550.
With this release, Anandtech's review mentioned that Intel is going to cut prices on the C2Q this month, probably to bring down the top chips down by about ~30$ or so, making the Q9550 the same price as the 940 and the Q9400 the same as the 920. Then, Intel would be ahead again at all stock clocks.
So, the release is good in a way. It made Intel think a bit and drop the prices on the C2Q, which is always a good thing for consumers. But the prices would have dropped more if the 940 had done a lot better.
With this release, Anandtech's review mentioned that Intel is going to cut prices on the C2Q this month, probably to bring down the top chips down by about ~30$ or so, making the Q9550 the same price as the 940 and the Q9400 the same as the 920. Then, Intel would be ahead again at all stock clocks.
So, the release is good in a way. It made Intel think a bit and drop the prices on the C2Q, which is always a good thing for consumers. But the prices would have dropped more if the 940 had done a lot better.
Uh, no they didn't do that in any of their recent CPU benchmark articles except on the last page of this one. Running a game at 640x480 is SPECIFICALLY testing the CPU and trying to avoid the limitations of the graphics card, I would damn well hope they tried to avoid other limitations... I hope. They do "real world" testing in other cases but this article wouldn't exactly fall into that definition, these were all canned benchmarks or component specific testing.
We usually know in advance which articles we will require flame suits and flak jackets for in the forums and in our E-Mails prior to these things being published.
That being said, I hadn't actually seen this article prior to it being published as I didn't have anything to do with it. I didn't expect Kyle's comments to be so "colorful."
- Joined
- May 18, 1997
- Messages
- 55,634
I cannot comment directly on the testing you are pointing to. I can only look at that, look at what we are doing, and tell you that the dynamics are not the same.
Trust me, we have been doing this long enough to know not to publish stories where systems are disk cacheing...unless its a story about disks or IO, or NOT enough RAM.
Another thing I wouldn't mind seeing is results from more games that people actually play. You've got FC2 and Crysis in there, but how about L4D, CoD 4/World at War, Fallout 3; you could even hark back to the HL2 days. I guarantee there are more people still playing HL2 than there are people playing Lost Planet and Quake 4 combined.
- Joined
- May 18, 1997
- Messages
- 55,634
- Joined
- May 18, 1997
- Messages
- 55,634
There was no real world testing done in this article. It is all benchmarks as specifically explained in the article.
The FarCry 2 page is a benchmark as well. A good benchmark, but a benchmark. I talked about it at length.
Well, the gist of that is to cover a wide swath of game engines. We could add those games, but do you think it is going to show you anything different? If you do, I would like to know why you think that. Also, with that comes the ability to do specifically repeatable benchmarks in a "scientific" fashion. I have not looked into Fallout 3, but would be interested in it.
Eh, I disagree with some choices in the article. But I don't really care about the Phenom II. I think it has it's place, I think Kyle doesn't recognize its place and that is about it. Different opinions, Kyle is just more... animated about his as usual!
- Joined
- May 18, 1997
- Messages
- 55,634
Still have that 4870x2 in my own box. It will make it over to my new ci7 build for sure.
I like the review, can I have a Phenom II as a prize? I could use it to put together a PC for my mom!
I heard they will be giving them away next week to clear room for the DDR3 version.
Please tell me what that place is in enthusiast terms. It sounds like you have an opinion worth sharing.
One thing I wish would have been at least touched on was compatibility with AM2 (not AM2+) motherboards. I remember AMD promising that AM3 chips would still work in AM2 sockets - any truth to that? If so, the Phenom II looks like it would be an excellent upgrade to replace my old X2 without needing to sell my soul to afford it....
- Joined
- May 18, 1997
- Messages
- 55,634
I still don't really see what that gets you. You can do cheap mobos with any type processor. I guess I just don't see being "extremely competitive" ie, not as fast but nearly, over a year after C2 launch.
No doubt you can build something that pulls up close, but why do that when you can something that will pull ahead for the same money?
- Joined
- May 18, 1997
- Messages
- 55,634
Please link your reliable resources so I can share in your exposure.
- Joined
- May 18, 1997
- Messages
- 55,634
Going to bed, have to get up in 4 hours to edit the next one.....
Please keep posting your questions, comments, and concerns. I will surely look at this thread tomorrow morning.
Please keep posting your questions, comments, and concerns. I will surely look at this thread tomorrow morning.
Alright guys, I averaged out a ton of benchmarks from different websites and here are the results:
At stock frequencies...
The Phenom II X4 940 3.0GHz is 0.83% faster on average than the Core 2 Quad Q9400 2.66GHz.
The Phenom II X4 940 3.0GHz is 5.50% slower on average than the Core 2 Quad Q9550 2.83GHz.
So the X4 940 is roughly equal to a Core 2 Quad Q9400, which is interesting for a few reasons:
The names are similar... X4 940 and Q9400. Coincidence?
The prices are the same... the X4 940 is $275 and the Q9400 is $270.
They both overclock to about 3.6Ghz on air... unfortunately for AMD this means an overclocked Q9400 will outperform an overclocked X4 940.
Hope that gives you guys a little something to think about.
At stock frequencies...
The Phenom II X4 940 3.0GHz is 0.83% faster on average than the Core 2 Quad Q9400 2.66GHz.
The Phenom II X4 940 3.0GHz is 5.50% slower on average than the Core 2 Quad Q9550 2.83GHz.
So the X4 940 is roughly equal to a Core 2 Quad Q9400, which is interesting for a few reasons:
The names are similar... X4 940 and Q9400. Coincidence?
The prices are the same... the X4 940 is $275 and the Q9400 is $270.
They both overclock to about 3.6Ghz on air... unfortunately for AMD this means an overclocked Q9400 will outperform an overclocked X4 940.
Hope that gives you guys a little something to think about.