AMD PHENOM 9550 vs Intel Core 2 Quad Q6600 – PCMark, Fritz Chess, and Super Pi Results
PCMark 7 and PCMark 8 are more of an overall system test. It uses many mixed workloads like web browsing, photo viewing, video playback, encoding, etc. The testing may rely on hard drive performance to some degree.
Fritz Chess Bench is very good for CPU Integer based testing. The program uses code based on the Fritz Chessbase program.
Super Pi XS 1.5 will calculate Pi to a specific number of digits and the time it takes to complete. The XS (Extreme Systems) version adds sub-milli-second timing to the results.
PCMark 7 includes system and individual tests and returns the results as a score. I only ran the basic test that is normally used for the demo version. Version 7 was originally designed for Windows 7. While it may work on other operating systems like Windows 8 and Windows 10 the results are not comparable to results from another OS.
| Processor↓ | Phenom 9550 Stock (2.20 GHz) is: | Phenom 9550 OC (2.64 GHz) is: | C2Q Q6600 Stock (2.40 GHz) is: | C2Q Q6600 OC (3.00 GHz) is: |
| Phenom 9550 (Stock 2.20 GHz) compared to: | Par | -9.06% | -6.04% | -16.97% |
| Phenom 9550 (OC 2.64 GHz) compared to: | +9.97% | Par | +3.32% | -8.68% |
| C2Q Q6600 (Stock 2.40 GHz) compared to: | +6.43% | -3.21% | Par | -11.62% |
| C2Q Q6600 (OC 3.00 GHz) compared to: | +20.43% | +9.51% | +13.15% | Par |
The Q6600 has the lead over the Ph9550. But stock for stock its only by 6.4%. The overclocked Ph9550 is 3% faster than the stock Q6600 while the overlocked Q6600 has around double digit wins over all.
| Processor↓ | Phenom 9550 @ 2.2 GHz is: | C2Q Q6600 @ 2.2 GHz is: |
| Phenom 9550 @ 2.20 GHz compared to: | Par (3DMark Score: 3,249) | PCMark: -0.46% |
| C2Q Q6600 @ 2.20 GHz compared to: | PCMark: +0.46% | Par (3DMark Score: 3,264) |
There was no real difference in this test. I used a different SSD than the main benchmark test due to the original SSD crashing as I was preparing the benchmarks for the 2.2 GHz Q6600. But even the Q6600 @ 2.4 GHz is still only showing +6.4% which has an extra 200 MHz in clock speed. So the numbers should be close.
PCMark 8 is an update to version 7 tailored for Windows 8. Like PCMark version 7 it is possible use on other operating systems as well, but just can’t compare the results from another OS.
| Processor↓ | Phenom 9550 Stock (2.20 GHz) is | Phenom 9550 OC (2.64 GHz) is | C2Q Q6600 Stock (2.40 GHz) is | C2Q Q6600 OC (3.00 GHz) is |
| Phenom 9550 (Stock 2.20 GHz) compared to: | Par | Score: -8.61% FPS: -8.30% | Score: -8.50% FPS: -11.75 | Score: -20.74% FPS: -28.41% |
| Phenom 9550 (OC 2.64 GHz) compared to: | Score: +9.42% FPS: +9.04% | Par | Score: +0.124% FPS: -3.77% | Score: -13.28% FPS: -21.94% |
| C2Q Q6600 (Stock 2.40 GHz) compared to: | Score: +9.28% FPS: +13.31% | Score: -0.124% FPS: +3.91% | Par | Score: -13.38% FPS: -18.88% |
| C2Q Q6600 (OC 3.00 GHz) compared to: | Score: +26.17% FPS: +31.70% | Score: +15.31% FPS: +28.11% | Score: +15.45% FPS: +23.28% | Par |
The results for version 8 increases the Q6600’s lead. The Q6600 is almost 10% higher than the Phenom 9550 and even the overclocked Phenom is slightly slower in Casual Gaming while the overclocked 6600 dominates all by 15-26% in the total score.
| Processor↓ | Phenom 9550 @ 2.2 GHz is: | C2Q Q6600 @ 2.2 GHz is: |
| Phenom 9550 @ 2.20 GHz compared to: | Par (3DMark Score: 2,208) Par (Gaming: 39.8) | PCMark:-5.03% Gaming FPS: -3.11% |
| C2Q Q6600 @ 2.20 GHz compared to: | PCMark:+5.30% Gaming FPS: +3.21% | Par (3DMark Score: 2,325) Par (Gaming: 41.08 fps) |
The Q6600 has a little more performance than the Ph9550 in version 8 than 7. About +5% in the PCMark score and 3% in the Casual Gaming score indicated in Frames Per Second.
Frirz Chess Bench is a good Integer test and returns the results in Kilo-nodes per second. You can specify how many threads to execute each run. This is a bench program that can give some insight on core scaling between the two processor designs. I ran the test from 4 threads down to a single thread. So we can see how a normal consumer style application may scale.
| Processor↓ | Phenom 9550 Stock (2.20 GHz) is | Phenom 9550 OC (2.64 GHz) is | C2Q Q6600 Stock (2.40 GHz) is | C2Q Q6600 OC (3.00 GHz) is |
| Phenom 9550 (Stock 2.20 GHz) compared to: | Par | 1 Core: -15.11% 2 Core: -15.56% 3 Core: -16.06% 4 Core: -14.55% | 1 Core: -24.73% 2 Core: -19.95% 3 Core: -19.70% 4 Core: -17.44% | 1 Core: -39.54% 2 Core: -35.50% 3 Core: -35.56% 4 Core: -34.62% |
| Phenom 9550 (OC 2.64 GHz) compared to: | 1 Core: +17.81% 2 Core: +18.42% 3 Core: +19.13% 4 Core: +17.02% | Par | 1 Core: -11.32% 2 Core: -5.20% 3 Core: -4.32% 4 Core: -3.39% | 1 Core: -28.77% 2 Core: -23.61% 3 Core: -23.23% 4 Core: -23.50% |
| C2Q Q6600 (Stock 2.40 GHz) compared to: | 1 Core: +32.86% 2 Core: +24.92% 3 Core: +24.52% 4 Core: +21.13% | 1 Core: +12.77% 2 Core: +5.48% 3 Core: +4.52% 4 Core: +3.51% | Par | 1 Core: -19.67% 2 Core: -19.42% 3 Core: -19.76% 4 Core: -20.80% |
| C2Q Q6600 (OC 3.00 GHz) compared to: | 1 Core: +65.40% 2 Core: +55.03% 3 Core: +55.18% 4 Core: +52.96% | 1 Core: +40.40% 2 Core: +30.90% 3 Core: +30.26% 4 Core: +30.70% | 1 Core: +24.50% 2 Core: +24.10% 3 Core: +24.62% 4 Core: +26.27% | Par |
The first take away is the huge advantage the “Conroe'” architecture has over the Phenom “Barcelona/Agena” architecture. This program leans heavily on Integer computations. The Core 2’s raw Integer performance is so much stronger. The stock Q6600 is actually 32% faster than the Ph9550 running a single thread. As the thread count increases, the Q6600’s advantage decreases a bit over the Phenom.
I broke down the results into two specific scaling examples in the table below. From a single thread to 3 threads and 1 thread to 4 threads.
| Processor↓ | 1 to 3 Core/Thread Scaling in % | 1 to 4 Core/Thread Scaling in % |
| Phenom 9550 (Stock 2.20 GHz) | +313%: | +410% |
| Phenom 9550 (OC 2.64 GHz) | +316% | +407% |
| C2Q Q6600 (Stock 2.40 GHz) | +293% | +374% |
| C2Q Q6600 (OC 3.00 GHz) | +293% | +380% |
The results have the Phenom scaling 313% going from 1 to 3 threads and 410% 1 to 4 threads. The stock Q6600 scales -7.8% less than the Phenom going from 1 to 3 threads. Going from 1 to 4 threads the overclocked Q6600 scaled 3% better than at stock likely due to the faster 1333 MHz FSB.
| Processor↓ | Phenom 9550 @ 2.2 GHz is: | C2Q Q6600 @ 2.2 GHz is: |
| Phenom 9550 @ 2.20 GHz compared to: | Par (1 Core: 1,269) Par (2 Core: 2,680) Par (3 Core: 3,972) Par (4 Core: 5,209) | 1 Core: -18.55% 2 Core: -12.93% 3 Core: -12.91% 4 Core: -11.24% |
| C2Q Q6600 @ 2.20 GHz compared to: | 1 Core: +22.77% 2 Core: +14.85% 3 Core: +14.82% 4 Core: +12.67% | Par (1 Core: 1,558) Par (2 Core: 3,078) Par (3 Core: 4,561) Par (4 Core: 5,869) |
Even clock for clock the Core 2 just walks all over the Phenom. Single threaded performance was over +22% higher than the Ph9550. Just like in the stock and overclocked tests, the Q6600’s performance had a notable drop off as you add more threads.
Note: The Q6600 Core performance scaling for 2 threads may rely more on how Windows does thread load balancing. Instead of running Thread 1 and Thread 2 on the 1st and 2nd Core (or 3rd and 4th), which is ideal since they are connected by an on-die crossbar, Windows could end up sending one of the threads to the 3rd or 4th Core which will require the cores to communicate across the FSB instead of the crossbar.
Super PI XS1.5 was a very popular tool for benchmarking CPU’s using x87 FPU instructions in the 90’s and 00’s. The program will calculate Pi to a certain number of digits after the decimal point using the Gauss-Legendre algorithm and time how long the execution takes to complete. This is a good raw FPU benchmark.
| Processor↓ | Phenom 9550 Stock (2.20 GHz) is | Phenom 9550 OC (2.64 GHz) is | C2Q Q6600 Stock (2.40 GHz) is | C2Q Q6600 OC (3.00 GHz) is |
| Phenom 9550 (Stock 2.20 GHz) compared to: | Par | 256 K: -19.41% 1 M: -18.13% 4 M: -17.16% | 256 K: -31.18% 1 M: -43.33% 4 M: -35.70% | 256 K: -83.14% 1 M: -100.34% 4 M: -86.71% |
| Phenom 9550 (OC 2.64 GHz) compared to: | 256 K: +16.26% 1 M:+15.34 % 4 M: +14.64% | Par | 256 K: -9.86% 1 M: -21.34% 4 M: -15.81% | 256 K: -53.37% 1 M: -69.60% 4 M: -59.36% |
| C2Q Q6600 (Stock 2.40 GHz) compared to: | 256 K: +23.77% 1 M: +30.23% 4 M: +26.30% | 256K: +8.97% 1 M: +17.58% 4 M: +13.65% | Par | 256 K: -39.60% 1 M: -39.77% 4 M: -37.60% |
| C2Q Q6600 (OC 3.00 GHz) compared to: | 256 KB: +45.40% 1 M: +50.08% 4 M: +46.44% | 256 KB:+34.80% 1 MB: +41.03% 4 MB: +37.25% | 256 K: +28.37% 1 M: +28.45% 4 M: +27.32% | Par |
I hope I got the percentage calculations correct since in this bench the lower the score the better. The Phenom again just doesn’t have a chance. The added clock speed of the Q6600 just compounds the results. The stock for stock results are showing the Q6600 up to 30% faster in 1 MB results. The overclocked Phenom is still -21% slower than the stock Q6600. The overclocked C2Q just doesn’t even see the Phenom in the rearview because it is so far ahead.
| Processor↓ | Phenom 9550 @ 2.2 GHz is: | C2Q Q6600 @ 2.2 GHz is: |
| Phenom 9550 @ 2.20 GHz compared to: | Par (256 KB: 6.52 sec) Par (1 MB: 34.86 sec) Par (4 MB: 181.60 sec) | 256KB: -33.33% 1MB: -46.41 4MB: -38.81% |
| C2Q Q6600 @ 2.20 GHz compared to: | 256 KB: +25.00% 1 MB: +31.70% 4 MB: +27.96% | Par (256 KB: 4.89 sec) Par (1 MB: 23.81 sec) Par (4 MB: 130.82 sec) |
The Core 2 is just outright superior to the Phenom even clock for clock. Intel has a 25% to 31% lead on the Phenom.