Original Link: https://www.anandtech.com/show/2837



i5 / P55 Lab Update -

We welcomed Anand back into the office with open arms this past weekend. He immediately started working on an in-depth analysis of clock for clock comparisons for several processors as a follow up to our Lynnfield launch article (among many other things). This analysis along with a quick i7/860 performance review will be available in the near future.

In the meantime, I have additional performance results using the P55 motherboard test suite along with some unusual results from our gaming selections. I am not going to dwell on with commentary in this short update. We will let the numbers speak for themselves at this point. Let’s get right to the results today, but first, the test setup.

Test Setup-



For our test results we setup each board as closely as possible in regards to memory timings and sub-timings. The P55 and 790FX motherboards utilized 8GB of DDR3, while the X58 platform contained 6GB. The P55 and X58 DDR3 timings were set to 7-7-7-20 1T at DDR3-1600 for the i7/920, i7/870, and i7/860 processors at both stock and overclocked CPU settings.

We used DDR3-1333 6-6-6-18 1T timings for the i5/750 stock setup as DDR3-1600 is not natively supported in current BIOS releases for this processor at a stock Bclk setting of 133. We had early BIOS releases that offered the native 1600 setting but stability was a serious problem and support was pulled for the time being. Performance is essentially the same between the two settings. When we overclocked the i5/750 to 3.8GHz, we utilized the same DDR3-1600 7-7-7-20 1T timings as the i7 setups.

The AMD 790FX setup is slightly different as trying to run DDR3-1600 at CAS 7 timings on the 1:4 divider is extremely difficult. DDR3-1600 is not natively supported on the Phenom II series so this divider is provided with a caveat that you are overclocking the memory bus. The same holds true for the Lynnfield (i7/8xx, i5/7xx) processors as DDR3-1333 is officially the highest memory speed supported and it is DDR3-1066 for the Bloomfield (i7/9xx).

Without resorting to some serious overvolting and relaxing of sub-timings, we set our AMD board up at DDR3-1600 8-8-8-20 1T timings. The difference in performance between C7 and C8 DDR3-1600 is practically immeasurable in applications and games on this platform. You might pick up an additional few tenths of second in SuperPi or a couple of extra points in AquaMark or 3DMark 2001SE, but otherwise performance is about equal.

However, in order to satisfy some of our more enthusiastic AMD supporters, we also increased our Northbridge speed from 2000MHz to 2200MHz to equalize, if not improve, our memory performance on the AMD system. Yes, we know, further increasing the NB speed will certainly result in additional performance but the focus of this short article is to show clock for clock results at like settings. Personally, I would run DDR3-1333 C6 with 8GB as this platform favors tighter timings over pure bandwidth.

Last, but not least, I only ran the i5/750 without turbo enabled and the P45/C2Q setup is missing. I am still completing those numbers. Anand will be providing additional analysis on the other Lynnfield processors in his update. The image gallery below contains our Everest memory results with each processor overclocked at similar memory settings along with voltage/uncore/subtiming options. I will go into these in more detail once the motherboard roundups start. For the time being, the 860/P55 offers slightly better throughput and latency numbers than the 920/X58 when overclocked. At stock, the numbers favor the Lynnfield, but primarily due to the turbo mode.

Other than that we are in a holding pattern on the P55 roundups at this time trying to figure out some unusual game and 3D Render results with our GTX275 video cards. I will discuss this problem in the game results.




Applications-

Cinema 4D R11 x64

Cinema 4D R11 is one of our favorite programs to create high-end 3D images and animations. This particular program is sensitive to memory bandwidth and is well threaded. We track the time it takes to render a swimming pool layout.

Application Performance - Maxon Cinema 4D R11 x64

 

Our i7/920 finishes first with the i7/860 just a second behind at 3.8GHz and with HT enabled. Disabling HT decreases performance 15% when these two CPUs are overclocked. The 965 BE is 27% slower at 3.8GHz than the i7/920 with HT enabled and 14% with HT disabled.

Turbo mode benefits the i5/750 greatly in this benchmark as turning it off results in the 750 being 16% slower. However, pure CPU speed allows the 965 BE to finish the benchmark about 10% quicker than the 750 with Turbo enabled and 23% with Turbo disabled with a 28% higher base clock speed. Of course, the 965 BE is 17% more expensive based on CPU cost comparisons.

LightWave 3D 9.6 x64

Another popular 3D rendering program is Lightwave 3D 9.6. In this test we time the rendering of a single frame from an office building animation. The time to render the full scene is approximately four and a half hours. This title is also well threaded and sensitive to both memory bandwidth and latency.

Application Performance - NewTek Lightwave 3D 9.6 x64

 

Lightwave also favors HT enabled processors with the 860/920 tying for first at 3.8GHz. Interestingly enough, Turbo mode on the i5/750 only offered a 3% improvement in this bench. The 965 BE scales very well as a 11% core speed increase nets a 10% improvement in the benchmark. However, the 965 BE still trails the 860/920 CPUs by 26% at 3.8GHz.

MainConcept Reference 1.61

One of our favorite video transcoding utilities is MainConcept Reference. We set our profile to iPOD HQ NTSC and then transcode a 651MB 1080P file to an iPOD HQ 34.7MB file.

Application Performance - MainConcept Reference 1.61

 

HyperThreading does not make a difference in this title with the 860/920 finishing in a dead heat again. Turbo mode for the i5/750 improves its scores by 11%. Once again, the 965 BE scales very well with an 11% increase in speed resulting in a 9% benchmark improvement. Otherwise, the 965 BE trails the 860/920/750 by 29% at 3.8GHz.

Sonar Producer 8 x64

We utilize Sonar Producer extensively at home when mixing various music tracks. This test performs a complex mix of five individual tracks into a single title. We covert these tracks into a WAV format utilizing a 192kHz sample rate along with all other options enabled. This title thrives on memory bandwidth.

Application Performance - Sonar Producer 8 x64

 

We finally see some separation between the 860 and 920 processors in this benchmark. Based on offline memory testing, we contribute this to the 920’s slightly better throughput under load conditions. The i5/750 performs 13% better with turbo enabled. The 965 BE performs 15% worse than the 750 at 3.8GHz.

Adobe Lightroom 2.4 x64

Lightroom is a quick and easy to use program for batch conversions on digital photos. We perform a standard conversion of 50 RAW images into the JPEG format.

Application Performance - Adobe Lightroom 2.4 x64

 

HyperThreading does not matter in this title. The 920 finishes just slightly ahead of the 860. Turbo mode makes 9% difference for the i5/750, just enough to place ahead of the 965 BE at stock speeds. Once overclocked, the 965 BE is 17% slower than the 750.

CyberLink MediaShow Espresso

We transcode a 370MB 1080AVCHD file into a HD friendly file suitable for publishing on YouTube. This program features GPU hardware accelerated decoding via ATI Stream or NVIDIA CUDA but is disabled in our test routines.

 

Application Performance - CyberLink Espresso

 

Even with a 740MHz advantage in clock speed, the 965 BE is just equal to the i5/750 with Turbo disabled in this application. Obviously, it appears this particular program behaves differently on Intel processors. Even at a 3.8GHz clock speed, the 965 BE trails the i5/750 by 24%. The 920/860 processors continue to be close, even though in this application the 920 has a slight advantage.



Multitasking-

The vast majority of our benchmarks are single task events that utilize anywhere from 23MB up to 1.4GB of memory space during the course of the benchmark. Obviously, this is not enough to fully stress test our 6GB or 8GB memory configurations. We devised a benchmark that would simulate a typical home workstation and consume as much of the 6GB/8GB as possible without crashing the machine.

We start by opening two instances of Internet Explorer 8.0 each with six tabs opened to flash intensive websites followed by Adobe Reader 9.1 with a rather large PDF document open, and iTunes 8 blaring the music selection of the day loudly. We then open two instances of Lightwave 3D 9.6 with our standard animation, Cinema 4D R11 with the benchmark scene, Microsoft Excel and Word 2007 with large documents, and finally Photoshop CS4 x64 with our test image.



Before we start the benchmark process, our idle state memory usage is 4.05GB. Sa-weet!



We wait two minutes for system activities to idle and then start playing Pinball Wizard via iTunes, start the render scene process in Cinema 4D R11, start a resize of our Photoshop image, and finally the render frame benchmark in Lightwave 3D. Our maximum memory usage during the benchmark is 5.62GB with 100% CPU utilization across all four or eight threads.

Application Performance - MultiTask Test - Lightwave 3D

 

Application Performance - MultiTask Test - Cinema 4D

 

Application Performance - MultiTask Test - Total Time

 

So far, our results have pretty much been a shampoo, rinse, and repeat event. I believe multitasking is what separates good systems from the not so good systems. I spend very little time using my system for gaming and when I do game, everything else is shutdown to maximize frame rates. Otherwise, I usually have a dozen or so browser windows open, music playing, several IM programs open and in use, Office apps, and various video/audio applications open in the background.

One or two of those primary applications are normally doing something simultaneously, especially when working. As such, I usually find this scenario to be one of the most demanding on a computer that is actually utilized for something besides trying to get a few benchmark sprints run before the LN2 pot goes dry.

The i5/750 results actually surprised me. The system never once felt “slow” but the results do not lie. The i5/750 had its head served on a platter at stock speeds, primarily due to the lack of Hyper-Threading when compared to the other choices. The 965 BE put up very respectable numbers and scaled linearly based on clock speed. An 11% increase in clock speed resulted in a 10% improvement in the total benchmark score for the 965 BE. You cannot ask for more than that.

At 3.8GHz clock speeds, it is once again a tossup between the 920 and 860 processors with HT enabled. The 920 did hold a slight advantage over the 860 at stock clock settings, attributable to slightly better data throughput when under load conditions. Otherwise, on the Intel side the i7/870 provided excellent results based on its aggressive turbo mode, although at a price.

Gaming-

We utilize the Ranch Small demo file along with the FarCry 2 benchmark utility. This particular demo offers a balance of both GPU and CPU performance.

Gaming Performance - Far Cry 2


We utilize FRAPS to capture our results in a very repeatable section of the game and report the median score of our five benchmark runs. H.A.W.X. responds well to memory bandwidth improvements and scales linearly with CPU and GPU clock increases.

Gaming Performance - H.A.W.X.


Your eyes are not deceiving you. After 100+ clean OS installs, countless video card, motherboard, memory and driver combinations, we have results that are not only repeatable, but appear to be valid. We also tracked in-game performance with FRAPS and had similar results. Put simply, unless we have something odd going on with driver optimizations, a BIOS bug, or a glitch in the OS, our NV cards perform better on the AMD platform than they do on the Intel platform. The pattern reverses itself when we utilize the AMD video cards.

It is items like this that make you lose hair and delay articles. Neither of which I can afford to have happen. However, we have several suppliers assisting us with the problem (if it is a problem) and hope to have an answer shortly. These results also repeat themselves in other games like H.A.W.X. and Left 4 Dead but not in Crysis Warhead or Dawn of War II. So, besides the gaming situation, we also see a similar pattern in AutoCad 2010 and other 3D rendering applications where GPU acceleration is utilized, it is just not as pronounced.



Quick Thoughts-

After a lot of speculation, it is obvious that the Core i7/860 and i7/920 platforms perform equally for the most part when compared on an equal clock basis, otherwise the 860 in Turbo mode is a better solution in most cases at stock speeds. The X58/920 combination will offer a very slight improvement in performance when the data pipeline is full; however it is very small, even under heavy multitasking conditions. The X58/920 does offer improved graphics performance in SLI or CF configurations.

While the 965 BE is a very competent processor, it is overshadowed at times by the performance of the turbo enabled i5/750. In the 3D rendering applications, the sheer clock speed advantage of the 965 BE certainly comes into play when comparing it directly to the i5/750. When comparing clock for clock speeds at 3.8GHz, the Intel processors offer a clear performance advantage over the 965 BE, especially in the well threaded applications that take advantage of Hyper-Threading. If you overclock your system and want the absolute best performance, the Intel processors used in our article today are your best choice.

Even though our initial benchmark selection is fairly limited, the overall pattern continues in the balance of our benchmark test suite, even in those applications specifically suggested by AMD. However, it is different story in several of our game benchmarks when the AMD system is paired with an Nvidia GTX275 (other GTX2xx variants as well). Why this is, we do not know yet given the results of the AMD HD 4890 on either platform.

And now for the Hot Computer Opinion (HCO). The X58/920 and 790FX/965BE platforms seemed slow in daily usage compared to the P55/860 setup or even the P55/750 at times. I am talking non-overclocked standard issue setups running a variety of applications, especially when multitasking. Of course, the difference is due to the Turbo modes employed on the Lynnfield processors.

While the perceived difference in performance is not as drastic as when moving from a hard drive to a decent SSD, it certainly is there. The performance benchmarks might tell another story at times, but if you just sit down and use a P55/860 platform and then move to an X58/920, 790FX/965BE, or P45/C2Q setup, the performance differences are noticeable in day to day usage. Even the P55/750 has its benefits and generally felt very “snappish” when under heavy loads.

Honestly, I really never thought I would say that after using a 790FX/965BE setup for several weeks and thinking afterwards I would have a very hard time recommending an X58/920 platform for typical home and gaming usage. When overclocked, the Intel Bloomfield/Lynnfield platforms basically performed equally on a clock for clock basis. The only differences were with the i5/750 in well-threaded applications. With HT enabled, the 860 and 920 are in a dead heat, except the Lynnfield platform will use about 70W less power for equal performance.

Unless you are a benchmark jockey, the dual x8 PCIe setup on the P55 is not going to be a performance hindrance with today’s video cards if you must run CrossFireX or SLI. Neither will the slightly better data throughput capabilities of the X58/920 when under heavy load conditions. I guess that really is the crux of the matter, unless you are a benchmark jockey then justifying a Bloomfield platform over a Lynnfield or even AMD’s Dragon platform is very difficult.

However, there are those that demand every last ounce of performance and the Bloomfield platform is the best choice for these particular users. I still really like the 790FX/965BE platform; in fact I would certainly purchase it over a P45/C2Q setup without question. When comparing it to the i5/750, the decision becomes more difficult, especially based on price.

However, considering my multitasking habits and the fact I do not overclock my work systems, the 965BE becomes the clear choice for me, until I compare it to the i7/860. Therein lies the problem, you can play the “what if” game all day and it will get you absolutely nowhere. In the end, you have to choose a platform that best suits your needs and budget. I just happen to think the clear choice for my particular needs is the i7/860 processor on a mid-range or even budget P55 motherboard. We will soon see why.

Log in

Don't have an account? Sign up now