Test Systems And Benchmarks
We used two test systems that share the same type of hard drive, CPU cooler, and memory in order to minimize the variables affecting performance. We also include results from a previously-tested Core i5-750 for reference, but only for application benchmarks, as the game benchmarks are affected by the newer graphic drivers.
To keep the CPUs running cool enough, we need aftermarket cooling capable of outperforming each company's stock cooler, yet cheap enough to apply to budget overclocking systems like the ones we're building. It doesn't make sense to put a $50 cooler on a $100 CPU, as the resulting cost would approach the price of the retail Core i5-750.
Once again, we chose Cooler Master's TX3. While it's not the most effective or quiet cooler out there, the TX3 does a much better job than a stock heatsink/fan combination, and the $20 price tag makes it an ideal fit for these budget CPUs. The Hyper TX3 works on both Socket AM3 and LGA 1156 interfaces, which makes for an ideal comparison in this case.
Now for the motherboards. For the AMD system, we chose Asus's M4A785TD-V EVO, a solid platform built around the 785G chipset that proved its memory support and overclocking prowess in our AMD 785G motherboard roundup. At $100 online, it is a fantastic low-cost AMD overclocking board.
Choosing a board for the Intel system is a little more complicated. The P55 chipset is our first choice, but most overclock-friendly P55 motherboards are a little more expensive, and we want to keep the budget low for this comparison. Since the Gigabyte H55M-USB3 performed very well in fellow editor Thomas Soderstrom's budget H55 chipset roundup, we'll take a leap of faith and give Gigabyte's H55M-UD2H a try. This board costs just under $90 online ($10 cheaper than our AMD board). On a side note, it's not often that the cost of an Intel CPU and motherboard is lower than a comparable AMD setup, but the LGA 1156 motherboard and CPU combo have a $10 advantage in this review.
| Intel Test System | AMD Test System | |||
|---|---|---|---|---|
| CPU | Intel Pentium G6950 2.8 GHz (Clarkdale) | AMD Phenom II X2 555 (Callisto), | ||
| Motherboard | Gigabyte H55M-UD2H LGA 1156 | Asus M4A785TD-V EVO Socket AM3 | ||
| Networking | Onboard Gigabit LAN controller | |||
| Memory | Mushkin PC3-10700 | Mushkin PC3-10700 | ||
| Graphics | Saphire Radeon HD 5850 | |||
| Hard Drive | Western Digital Caviar Black 640GB | |||
| Power | Corsair CMPSU-750HX 750W | |||
| Software and Drivers | ||||
| Operating System | Microsoft Windows 7 x64 | |||
| DirectX version | DirectX 11 | |||
| Graphics Drivers | ATI Catalyst 10.3 | |||
| Benchmark Configuration | |
|---|---|
| 3D Games | |
| Crysis | Patch 1.2.1, DirectX 10, 64-bit executable, benchmark tool High Quality, No AA |
| Fallout 3 | Patch 1.7, Saved Game "Capital Wasteland" (60 sec) Highest Details, No AA, No AF |
| Far Cry 2 | Patch 1.03, DirectX 10, in-game benchmark Ultra High Quality, No AA |
| Tom Clancy's H.A.W.X | Patch 1.02, DirectX 10.1, in-game benchmark Highest Settings, No AA |
| World in Conflict | Patch 1009, DirectX 10, timedemo Very High Details, No AA / No AF |
| Audio/Video Encoding | |
| iTunes 8 | Version: 8.2.1.6 (x64) Audio CD ("Terminator II" SE), 53 min Default format AAC |
| Lame MP3 | Version: 3.98.2, wave to MP3 Audio CD "Terminator II" SE, 53 min wave to MP3 |
| TMPGEnc 4.0 Express | Version: 4.7.3.292 Import File: "Terminator 2" SE DVD (5 Minutes) Resolution: 720x576 (PAL) 16:9 |
| DivX 6.8.5 | Encoding mode: Insane Quality Enhanced multithreading enabled using SSE4 Quarter-pixel search |
| XviD 1.2.2 | Display encoding status = off |
| MainConcept Reference 1.6.1 Reference H.264 Plugin Pro 1.5.1 | MPEG2 to MPEG2 (H.264), MainConcept H.264/AVC Codec, 28 sec HDTV 1920x1080 (MPEG2), Audio: MPEG2 (44.1 KHz, 2 Channel, 16-Bit, 224 Kb/s), Mode: PAL (25 FPS) |
| Productivity | |
| Adobe Photoshop CS4 (64-bit) | Version: 11.0 Extended, Filter 15.7MB TIF Image Radial Blur, Shape Blur, Median, Polar Coordinates |
| Autodesk 3ds Max 2010 | Version: 11.0, Rendering Dragon Image at 1920x1080 (HDTV) |
| WinRAR 3.90 | Version x64 3.90, Dictionary = 4,096KB, Benchmark: THG-Workload (334MB) |
| WinZip 12 | Version 12.1, WinZip Command Line Version 3.0 Compression = Best, Benchmark: THG-Workload (334MB) |
| Synthetic Benchmarks | |
| 3DMark Vantage | Version: 1.01, GPU and CPU scores |
| PCMark Vantage | Version: 1.00, System, Memory, Hard Disk Drive benchmarks, Windows Media Player 10.00.00.3646 |
| SiSoftware Sandra 2009 SP4a | Version 2009.9.15.130, CPU Test = CPU Arithmetic / MultiMedia, Memory Test = Bandwidth Benchmark |
I would love to see a comparison between i3-530 and X3 435 with GTA4.
This is exactly what I've been waiting for. The numbers matched what I had already assumed. I got the 555 but I will consider the intel next time to change it up.
One thing to consider is the Pentium G6950 is tied to a crippled platform, whereas the Phenom II can be used with an 890FX, which has more PCI-E lanes for Crossfire, comes with SATA 6Gbps, and can have USB 3.0 added without either running degraded, or using PCI-E lanes used for the video card.
The AMD platform gives you more choices (integrated graphics, discreet graphics in several flavors, a lot of PCI lanes, or a few), and an unlocked multiplier.
All these are important considerations.
Considering you fryed one of the intel cpus quite quickly with only 7% more voltage...
Id like to see a serious stability test on both cpus. A couple days with a graphic benchmark on loop as well as prime95 running an instance on each core would do it.
I dont understand they talk about the dormant cores and you may not be successful, this is a dual core shoot out, so you intended buying a dual core, why not get the phenom II 555 and see if the cores unlock if they dont well no biggie still a fast CPU but if they do BONUS, also they dont tell you that with 2 cores the phenom can easily get OVER 4GHZ, this is typical of toms not putting everything into the tests, definately Intel fans
One thing to consider is the Pentium G6950 is tied to a crippled platform, whereas the Phenom II can be used with an 890FX, which has more PCI-E lanes for Crossfire, comes with SATA 6Gbps, and can have USB 3.0 added without either running degraded, or using PCI-E lanes used for the video card.
The AMD platform gives you more choices (integrated graphics, discreet graphics in several flavors, a lot of PCI lanes, or a few), and an unlocked multiplier.
All these are important considerations.
I dont understand they talk about the dormant cores and you may not be successful, this is a dual core shoot out, so you intended buying a dual core, why not get the phenom II 555 and see if the cores unlock if they dont well no biggie still a fast CPU but if they do BONUS, also they dont tell you that with 2 cores the phenom can easily get OVER 4GHZ, this is typical of toms not putting everything into the tests, definately Intel fans
I don't know if this is ROFL or just sad ? Try reading the article. Your embarrassing yourself.
More faulty logic by AMD fanboys. Which is it ? A budget bang for your buck rig,
H55/Clarksdale=200 dollars
or
890FX ($160.00 MIN)+ 555=260, all so you can buy another cpu, next year, that does not exist yet ?
and 890fx, you HAVE to buy a DISCRETE graphics card now.
AMD will love you , if you invest in all of this hardware , with plans to buy more, lol.
Thanks Don ... another solid article without the fanboi slant.
I don't know of any business who is better off with an Intel CPU these days...
Businesses do not overclock and the AMD Platform with an ATI4200 onboard is just so much better then anything intel has to offer... Encoding ? Use and AthlonX4...
The only good thing about the Pentium would be trying to break overclocking records... other then that... it's just a cheep cpu that fails against any amd tricore.
I don't know if this is ROFL or just sad ? Try reading the article. Your embarrassing yourself.More faulty logic by AMD fanboys. Which is it ? A budget bang for your buck rig,H55/Clarksdale=200 dollarsor890FX ($160.00 MIN)+ 555=260, all so you can buy another cpu, next year, that does not exist yet ?and 890fx, you HAVE to buy a DISCRETE graphics card now.AMD will love you , if you invest in all of this hardware , with plans to buy more, lol.
Hmmmm, talk about embarrassing yourself - didn't you even bother to find out if your numbers were right before posting? You can get a 890FX for $140, not $160 MIN as you stated. For $155 you can get one with USB 3.0 and SATA 6Gbps, and it's not implemented with the compromises inherent with the LGA 1156 platform.
You're quite incorrect about needing a discreet GPU. AMD sells the 890GX, 790GX, 785G, 760G, and 880G. In fact, the platform they used had an integrated GPU. The nice thing with the AMD platform is, they have sideport memory, so you don't degrade CPU performance when you use the IGP due to memory contention.
So, I can get the AMD platform with motherboards around $60 with an IGP, or I can get a powerful platform with two real PCI-E 16x slots, USB 3.0, and SATA 6.0 Gbps for $155. You don't have the same choices with the Pentium G6950 platform in either direction. AM3 processors have a very diverse selection of platform.
So, is your contention that choice is bad?
You know they do this all the time... When they start comparing the game performance, they drop the i5 from the comparison charts. WHY??????
Very interesting article.
C00lit's point about businesses not overclocking is good, although I'm not sure strictly business users come to Tom's for guidance.
I think these are both lab chips; maybe fun to test and tweak and play with, but far from the most sensible choice for real use. The AMD X3 chips look like fairly definitive bang for buck winners all around.
You know they do this all the time... When they start comparing the game performance, they drop the i5 from the comparison charts. WHY??????
I mention this in the article. The new graphics drivers really changed the game results compared to the old ones, and we don't have the i5 system here to retest. Since the game results weren't comparable anymore we left the i5 numbers out of gaming results.
I dont understand they talk about the dormant cores and you may not be successful, this is a dual core shoot out, so you intended buying a dual core, why not get the phenom II 555 and see if the cores unlock if they dont well no biggie still a fast CPU but if they do BONUS, also they dont tell you that with 2 cores the phenom can easily get OVER 4GHZ, this is typical of toms not putting everything into the tests, definately Intel fans
I don't quite follow your logic. Are you being sarcastic? They did try to unlock the cores and they were successful. bonus. They also did get two CPU cores to 4+ GHz...
So they did put everything into the test that you are accusing them of not putting into the test...
My take on the drop of i5 is that their numbers for i5 were based on systems using older drivers, and there was not enough time to re-run them all with updated drivers. I have no complaints about this; I don't think any rational individual was expecting this to be a be-all, end-all article on CPU performance. I5 was not one of the CPUs being compared anyway, and confounding the results with multiple driver versions would just bring out the more trolls.
all the graphs are jumbled. can someone arrange them in a high/low setup?
My only disagreement was with the final word, "At the end of the day, both of these CPUs offer solid value." No, they don't; not when there are clearly better bang/buck choices available from both AMD and Intel.
all the graphs are jumbled. can someone arrange them in a high/low setup?
If they went high-to-low they'd simply be jumbled in a different way. This way, the relative position of each CPU is in the same place in each chart.
My take on the drop of i5 is that their numbers for i5 were based on systems using older drivers, and there was not enough time to re-run them all with updated drivers. I have no complaints about this; I don't think any rational individual was expecting this to be a be-all, end-all article on CPU performance. I5 was not one of the CPUs being compared anyway, and confounding the results with multiple driver versions would just bring out the more trolls.
I agree with you about the i5, but I do think article comparing the i3, Pentium, Athlon II x2 260, and Pentium G6950 would be kind of informative.
As much as people like to see expensive processors, the reality is that the lower end processors are often what people buy. In particular, I think the i3 is probably worth the extra money, considering it not only gets you 1 MB cache, but also hyper-threading which can be useful.
Also, I saw a review on another site on the Athlon II x2 260, which, when compared to the 255, is much faster than it should be. Performance went up more than the clock speed, and when I questioned them why, they brushed it off and just said they were only aware it was a faster clock speed. Maybe they tested wrong, maybe there is no difference, but another review showed the difference between the 260 and the Phenom II 255 to be quite small as well.
I think a lot people would be curious about matching low-end platforms on a cost basis. For example, if I save $25 on the processor, and spend it on the video card, what do the results look like. I would be impossible to do in one article, but I think a series of articles would be very interesting, covering different price points. It's time consuming, but, I think it's got broad interest and would get page hits. Of course, I could be wrong.
I was looking around yesterday for benchmarks on the new 45W AMD chips to see if one of them would be "acceptable," even though I'd expect them to be somewhat slow. I was quite suprised to see on quite a few tests (including games) they beat an i3, sometimes severely (others they lost, equally severely).
Just checked History, it was here: http://www.cpu-world.com/CPUs/K10/ [...] IBOX).html
and this was actually the 405e, not the new 415e or 6x5e.