|
AMD Athlon II X4 645 3.10 GHz
Introductionhttp://www.techpowerup.com/reviews/A...mages/logo.jpg http://www.techpowerup.com/reviews/A...images/amd.jpg First of all, I would like to thank AMD for providing the tested AMD Athlon II X4 645 processor sample, MSI for providing the 890FXA-GD70 motherboard and Links for providing the rest of the test setup. After few months of keeping a low profile, AMD is refreshing its whole processor retail offer from top to bottom. While introducing new and faster models with increased clock, AMD is keeping its top price bars, lowering the prices of existing models and introducing new, faster ones in their place, like the AMD Athlon II X4 645. In this review we take a look at the new addition to Athlon II family models, the fastest quad core Athlon II yet, named Athlon II X4 645 and clocked at an impressive 3.10 GHz, considering this is to be a budget processor. Aiming to be a powerful yet affordable general purpose processor, the new Athlon II X4 645 with its faster clocked cores is stepping into the upper middle market segment, ruled by slower Phenom II models and Intel's Core i3 and i5 processors. It will be interesting to see if the Athlon II X4 Series matured enough to take on these new challenges. Packaging and ContentsWe received our test sample in tray package, but the retail version packaging comes with AMD certified cooler, manual and three year warranty. Athlon II X4 645http://www.techpowerup.com/reviews/A...ront_small.jpg http://www.techpowerup.com/reviews/A...back_small.jpg Long after retiring the Phenom II X4 800 series, AMD is introducing a new member to the forgotten family, the Phenom II X4 840 quad core processor. New model comes with quite high clock frequencies and very low price tag. Low enough to raise your eyebrows in disbelief. AMD is asking just $102 for quad core Phenom II processor, and if some of you can remember, that was the starting price for 2.60 GHz Athlon II X4 620 introduced a year ago, so one would think AMD came a long way, and one would be somewhat deceived. What AMD has done here is mix up the Athlon II quad Propus core with Phenom II brand. After introduction of Athlon II X4 645, quad core Athlon II models clocking above 3.0 GHz were starting to catch up to slower Phenom II models, so AMD decided to upgrade the low cost Propus core to Phenom II brand. In other words, what we have here is Athlon II X4 "650" renamed to Phenom II X4 800 series. In terms of specifications, it is identical to every other Athlon II X4 SKU, and the one thing that differed Athlon II from Phenom II series, L3 cache buffer is no more. The older Phenom II X4 800 series of processors had cut down L3 cache, to 4 MB. The new Phenom II X4 800 series, starting with 840 model we have here, has no L3 cache at all. Not a big surprise since it's using Propus core, but it comes as a surprise that AMD decided to take this step, and confuse its buyers even more. But then again, for just over $100 there's no room for complaints, right? Either way, what we have here is a budget quad core processor, based on Propus core used in Athlon II X4 and X3 series, and clocked to 3.20 GHz. Phenom II X4 840 is based on C3 revision of Propus core, just like the Athlon II X4 645 we reviewed recently, so the benefits of lower power consumption and lower heat output are present here as well. The processor's frequency multiplier is locked to x16, and only adjustable downwards, so the only way to overclock is via raising the HTT bus. Core voltage is set a bit higher than Phenom II models, to 1.40 V, but TDP is still kept to acceptable 95W. Like any other Phenom II or Athlon II model, backwards compatibility with AM2+ boards and DDR2 memory is here, and very useful considering the market this processor is targeting. Since we covered all of the architectural features of AMD's Propus core in our first encounter with Athlon II X4, I would like to redirect you to the X4 620 spec page of the review if you're interested in more details about the processor core itself (LINK). http://www.techpowerup.com/reviews/A...images/die.jpg http://www.techpowerup.com/reviews/A...ages/stock.jpg Test Systems<table border="1" cellpadding="3" cellspacing="0" class="ramtable" width="450"> <tr align="center"> <th colspan="2" scope="row" style="font-size:larger;text-align:center">Test System</th> </tr> <tr> <th scope="row">Motherboard:</th> <td scope="row">AMD: ASUS M4A79T Deluxe<br />MSI 890FXA-GD70<br />ASUS Crosshair IV Formula<br />Intel LGA1366: ASUS P6T<br />Intel LGA1156: ASUS P7P55D PRO<br />Intel LGA775: ASRock P45X3 Deluxe</td> </tr> <tr> <th scope="row">Cooler:</th> <td scope="row">Scythe Katana III</td> </tr> <tr> <th scope="row">Memory:</th> <td scope="row">2 x 2048 MB G.Skill Perfect Storm PC-16000 DDR3<br />@ 1333 MHz 7-7-7<br />@ 1066 MHz 7-7-7 for C2D E7000 </tr> <tr> <th scope="row">Graphic card</th> <td valign="top" scope="row">VTX ATI Radeon HD 5850</td> </tr> <tr> <th scope="row">Harddisk:</th> <td valign="top" scope="row">Samsung HD161HJ 160 GB</td> </tr> <tr> <th scope="row">Power Supply:</th> <td valign="top" scope="row">Enermax Liberty 620 W</td> </tr> <tr> <th scope="row">Software:</th> <td valign="top" scope="row">Windows 7 Ultimate 64-bit</td> </tr> <tr> <th scope="row">Drivers:</th> <td valign="top" scope="row">Catalyst 9.12 Hotfix</td> </tr> </table> Below you can see the entire list of programs used to test our processors. Most of the tests make good use of multithreading, but just like real-life situations, there are some tests that prefer higher clock frequencies over number of cores or size of L2/L3 cache buffer. Since Windows 7 64-bit is used from now on as base for processor benchmarks, some of benchmarks are used in 64-bit versions as well. Benchmark Suite
SpecificationsIn the table below you can review detailed specifications of AMD Phenom II X6 1055T and other models that will show up later in the performance graphs. <table border="1" class="resulttable" cellspacing="0" cellpadding="3"> <tr> <th>Intel Processors</th> <td>Intel <br />Core 2 Duo<br />E7500</td> <td>Intel <br />Core 2 Duo<br />E7600</td> <td>Intel <br />Core 2 Duo<br />E8400</td> <td>Intel <br />Core 2 Quad<br />Q8300</td> <td>Intel <br />Core 2 Quad<br />Q8400</td> <td>Intel <br />Core i3 <br />530</td> <td>Intel <br />Core i3 <br />540</td> <td>Intel <br />Core i5 <br />661</td> </tr> <tr> <th>Number of cores</th> <td align="right">Dual</td> <td align="right">Dual</td> <td align="right">Dual</td> <td align="right">Quad</td> <td align="right">Quad</td> <td align="right">Dual</td> <td align="right">Dual</td> <td align="right">Dual</td> </tr> <tr> <th>Core codename</th> <td align="right">Wolfdale 4M</td> <td align="right">Wolfdale 4M</td> <td align="right">Wolfdale</td> <td align="right">Yorkfield 4M</td> <td align="right">Yorkfield 4M</td> <td align="right">Clarkdale</td> <td align="right">Clarkdale</td> <td align="right"> Clarkdale</td> </tr> <tr> <th>Socket</th> <td align="right">LGA 775</td> <td align="right">LGA 775</td> <td align="right">LGA 775</td> <td align="right">LGA 775</td> <td align="right">LGA 775</td> <td align="right">LGA 1156</td> <td align="right">LGA 1156</td> <td align="right">LGA 1156</td> </tr> <tr> <th>Core speed</th> <td align="right">2930 MHz</td> <td align="right">3060 MHz</td> <td align="right">3000 MHz</td> <td align="right">2500 MHz</td> <td align="right">2660 MHz</td> <td align="right">2930 MHz</td> <td align="right">3060 MHz</td> <td align="right">3330 MHz</td> </tr> <th>L2 Cache</th> <td align="right">3 MB</td> <td align="right">3 MB</td> <td align="right">6 MB</td> <td align="right">4 MB</td> <td align="right">4 MB</td> <td align="right">256 KB/core</td> <td align="right">256 KB/core</td> <td align="right">256 KB/core</td> </tr> <th>L3 Cache</th> <td align="right">-</td> <td align="right">-</td> <td align="right">-</td> <td align="right">-</td> <td align="right">-</td> <td align="right">4 MB</td> <td align="right">4 MB</td> <td align="right">4 MB</td> </tr> <th>Process Node</th> <td align="right">45 nm</td> <td align="right">45 nm</td> <td align="right">45 nm</td> <td align="right">45 nm</td> <td align="right">45 nm</td> <td align="right">32/45 nm</td> <td align="right">32/45 nm</td> <td align="right">32/45 nm</td> </tr> <th>Core die size</th> <td align="right">82 mm²</td> <td align="right">82 mm²</td> <td align="right">107 mm²</td> <td align="right">164 mm²</td> <td align="right">164 mm²</td> <td align="right">81+114 mm²</td> <td align="right">81+114 mm²</td> <td align="right">81+114 mm²</td> </tr> <th>TDP</th> <td align="right">65 W</td> <td align="right">65 W</td> <td align="right">65 W</td> <td align="right">95 W</td> <td align="right">95 W</td> <td align="right">73 W</td> <td align="right">73 W</td> <td align="right">87 W</td> </tr><tr> <th>Price</th> <td align="right">$ 124</td> <td align="right">$ 150</td> <td align="right">$ 170</td> <td align="right">$ 170</td> <td align="right">$ 175</td> <td align="right">$ 115</td> <td align="right">$ 125</td> <td align="right">$ 210</td> </tr> </table> <table border="1" class="resulttable" cellspacing="0" cellpadding="3"> <tr> <th>AMD Processors</th> <td>AMD <br />Phenom II <br />X2 550 BE</td> <td>AMD <br />Athlon II<br />X4 630</td> <td><strong>AMD <br />Athlon II<br />X4 645</strong></td> <td>AMD <br />Phenom II <br />X3 720 BE</td> <td>AMD <br />Phenom II <br />X4 925</td> <td>AMD <br />Phenom II <br />X4 945</td> </tr> <tr> <th>Number of cores</th> <td align="right">Dual</td> <td align="right">Quad</td> <td align="right"><strong>Quad</strong></td> <td align="right">Triple</td> <td align="right">Quad</td> <td align="right">Quad</td> </tr> <tr> <th>Core codename</th> <td align="right">Callisto</td> <td align="right">Propus</td> <td align="right"><strong>Propus</strong></td> <td align="right">Heka</td> <td align="right">Deneb</td> <td align="right">Deneb</td> </tr> <tr> <th>Socket</th> <td align="right">AM2+/AM3</td> <td align="right">AM2+/AM3</td> <td align="right"><strong>AM2+/AM3</strong></td> <td align="right">AM2+/AM3</td> <td align="right">AM2+/AM3</td> <td align="right">AM2+/AM3</td> </tr> <tr> <th>Core speed</th> <td align="right">3100 MHz</td> <td align="right">2800 MHz</td> <td align="right"><strong>3100 MHz</strong></td> <td align="right">2800 MHz</td> <td align="right">2800 MHz</td> <td align="right">3000 MHz</td> </tr> <th>L2 Cache</th> <td align="right">512 KB/core</td> <td align="right">512 KB/core</td> <td align="right"><strong>512 KB/core</strong></td> <td align="right">512 KB/core</td> <td align="right">512 KB/core</td> <td align="right">512 KB/core</td> </tr> <th>L3 Cache</th> <td align="right">6 MB</td> <td align="right">-</td> <td align="right"><strong>-</strong></td> <td align="right">6 MB</td> <td align="right">6 MB</td> <td align="right">6 MB</td> </tr> <th>Process Node</th> <td align="right">45 nm</td> <td align="right">45 nm</td> <td align="right"><strong>45 nm</strong></td> <td align="right">45 nm</td> <td align="right">45 nm</td> <td align="right">45 nm</td> </tr> <th>Core die size</th> <td align="right">258 mm²</td> <td align="right">168 mm²</td> <td align="right"><strong>168 mm²</strong></td> <td align="right">258 mm²</td> <td align="right">258 mm²</td> <td align="right">258 mm²</td> </tr> <th>TDP</th> <td align="right">80 W</td> <td align="right">95 W</td> <td align="right"><strong>95 W</strong></td> <td align="right">95 W</td> <td align="right">95 W</td> <td align="right">125 W</td> </tr><tr> <th>Price</th> <td align="right">$ 91</td> <td align="right">$ 96</td> <td align="right"><strong>$ 122</strong></td> <td align="right">-</td> <td align="right">-</td> <td align="right">$ 136</td> </tr> </table> Everest UltimateEverest is great for quick diagnostic of computer components and for checking some of the CPU's basic synthetic performance, as well as memory bandwidth and latency. Everest Ultimate version 530 is used to check the Memory Read/Write/Copy/Latency performance. http://www.techpowerup.com/reviews/A...erest_read.gif http://www.techpowerup.com/reviews/A...rest_write.gif http://www.techpowerup.com/reviews/A...erest_copy.gif http://www.techpowerup.com/reviews/A...st_latency.gif Synthetic BenchmarksPossibly the best known and widely used synthetic benchmark, 3D Mark Vantage tests the gaming performance of the PC. PC Mark Vantage on the other hand tests other performance aspects of the PC like video encoding, home entertainment multimedia features, working in an office environment, and basic gaming. 3D Mark Vantage v1.0.1.http://www.techpowerup.com/reviews/A...3dmark_cpu.gif http://www.techpowerup.com/reviews/A...mark_total.gif PC Mark Vantage v1.0.1.http://www.techpowerup.com/reviews/A.../pcvantage.gif wPrime 1024M and SuperPI 1MwPrime is a benchmarking application designed to use a highly multithreaded approach to calculate the square-roots of large amounts of numbers, while SuperPI calculates pi number to a specified number of digits after the decimal point. Both applications are primarily used by overclockers to check the stability and performance of overclocked CPUs. It's clear from results that wPrime is a multithreaded benchmark that uses all of the cores no matter if they are simulated with HyperThreading, while SuperPI uses just one core. http://www.techpowerup.com/reviews/A...ges/wprime.gif http://www.techpowerup.com/reviews/A...es/superpi.gif ImToo Audio Encoder v2.1.78ImToo Audio Encoder is representing commercial versions of audio encoders. It's easy to use, with broad audio format support, and for this test it is used to encode a high quality WMV file into 128 kbs quality MP3 format, also using Lame encoder. http://www.techpowerup.com/reviews/A...ages/imtoo.gif x264 HD Benchmark v3.0x264 HD Benchmark tests the processor's performance when converting a small 720p HD movie clip into a x264 file. It runs the encoding process four times with two passes on each encoding run and returns results in frames per second (FPS). Since video encoding is done in frames, higher frame rate translates in lower time needed to finish encoding, and in the results table you'll find average results of all eight encoding runs. http://www.techpowerup.com/reviews/A...mages/x264.gif Handbrake 0.9.4Handbrake video encoding utility is small, free, easy to use and works great when encoding DVDs to MKV files. For this test the first three chapters of the "U2 Go Home" DVD are encoded to MKV format using the Regular - High preset. It uses x264 codec with AAC audio codec. Needed time in seconds is measured, how long it takes the CPU to encode the video. As you can see from the graphs it scales almost perfectly with multithreaded CPUs. http://www.techpowerup.com/reviews/A.../handbrake.gif Xilisoft Video Converter Ultimate v5.1.26Xilisoft Video Converter Ultimate is great when you need to do a lot of day-to-day small video encoding from one format to another. It has an extremely wide range of encoding presets and supports almost every known video format. For this test the HD H.264/MPEG4 AVC preset is used to encode a 1080p high definition movie trailer clip. http://www.techpowerup.com/reviews/A...s/xilisoft.gif DivX Converter 7.2A long time ago DivX was the format almost everyone used every day and now, finally it’s getting the x264 codec with its 7.2 version. Too little, too late and with bad multithreading support, but still used by a wide range of users. http://www.techpowerup.com/reviews/A...mages/divx.gif Photoshop CS4 Retouch Artists Speed TestA must-have tool for all photographers and graphics designers, Adobe Photoshop allows you to create wonders out of poorly taken photos. Retouch Artist Speed Test is used to test performance of processors. A small Action Script does a lot of layer and effect work on test image and it manages to produce some good multithreading load on processor cores. http://www.techpowerup.com/reviews/A...toshop_cs4.gif Cinebench R10 64bit3D rendering tools are also used to test CPU performance. Cinebench R10, Blender and POV Ray simulate real world performance with integrated benchmark scripts. 3D rendering has been Intel's home ground for a long time, but AMD is making progress with its new models and is not that far behind Intel's processors. http://www.techpowerup.com/reviews/A.../cinebench.gif Blender 2.49http://www.techpowerup.com/reviews/A...es/blender.gif POV Ray 3.7b 64bithttp://www.techpowerup.com/reviews/A...es/pov_ray.gif File Compression7Zip is a free archiving tool and strong alternative to popular WinRAR. Integrated benchmark scripts are used in both 7Zip and WinRAR to test the performance of CPUs with multithreading enabled. One real life test with WinRAR is simulated while measuring the time needed to compress a ~1 GB folder with a few hundred smaller files. We can see that multithreading support does not work as well as it does in pure benchmark scripts. 7Zip 32MBhttp://www.techpowerup.com/reviews/A...mages/7zip.gif WinRAR 3.9 Benchmarkhttp://www.techpowerup.com/reviews/A...ges/winrar.gif WinRAR 3.9 compressingWhile both Core i3 and Core i5 661 processors do okay in synthetic archive benchmarks, in real life compression there is a issue with Intel's Speed Step and hence much weaker results than from the competition. Older CPUs are benchmarked with Speed Step off for result consistency, but with newer Intel processors if you were to disable Speed Step on most models you automatically lose Turbo Boost features and with that a few percent in almost every application, so they are tested with Turbo Boost always set to On status. What happens with WinRAR compression is that Intel's Turbo Boost fails to speed up the cores to the maximum value, and they get stuck somewhere between low idle and full load speed. Disabling Turbo Boost would improve WinRAR real-life compression performance but would also increase idle power consumption and decrease general performance in other applications. http://www.techpowerup.com/reviews/A.../winrar _c.gif Gaming TestsAnd finally some gaming benchmarks, the favorite review page for all gamers. Five popular titles are chosen to represent every game genre. Crysis and Modern Warfare for FPS gamers, Resident Evil 5 for adventure gamers, GTR Evolution and DiRT 2 for races and Dawn of War 2 for those who like to play RTS genre games. In-game quality details for all resolutions are set to following:
Crysis Warhead v1.2 64bithttp://www.techpowerup.com/reviews/A...rysis_1024.gif http://www.techpowerup.com/reviews/A...rysis_1680.gif Modern Warfare 2http://www.techpowerup.com/reviews/A...s/mw2_1024.gif http://www.techpowerup.com/reviews/A...s/mw2_1680.gif Warhammer II v1.5http://www.techpowerup.com/reviews/A.../dow2_1024.gif http://www.techpowerup.com/reviews/A.../dow2_1680.gif DiRT2http://www.techpowerup.com/reviews/A...dirt2_1024.gif http://www.techpowerup.com/reviews/A...dirt2_1680.gif GTR Race ON!http://www.techpowerup.com/reviews/A...s/gtr_1024.gif http://www.techpowerup.com/reviews/A...s/gtr_1680.gif Resident Evil 5 benchmarkhttp://www.techpowerup.com/reviews/A...vil_5_1024.gif http://www.techpowerup.com/reviews/A...vil_5_1680.gif System Power ConsumptionThe four result values are as following:
When first released, C'n'C and Speed Step were not so well accepted by desktop users because many bugs in performance and overclocking stability appeared when activated. But since then those technologies have greatly evolved. Today it is almost unthinkable to turn off Speed Step/C1E with Intel Core ix processors because you would automatically lose Turbo Boost benefits, and there is an increasing number of enthusiasts who overclock their processors and then push for additional MHz using Turbo Boost. So our measurements of Idle and Load CPU states will shift to those with power management features turned ON. Today idle power usage can have a noticeable impact on your electricity bill if you keep your computer on most of the day, and most people do just that. So a bad implementation of power management technologies is and will be considered a product flaw. http://www.techpowerup.com/reviews/A.../power_cnc.gif http://www.techpowerup.com/reviews/A...power_idle.gif http://www.techpowerup.com/reviews/A...power_load.gif Overclockinghttp://www.techpowerup.com/reviews/A.../images/oc.jpg Overclocking results for Athlon II X4 645 were unimpressive to say the least. Just over 3.70 GHz is not something you'd want to brag about these days, but keep in mind that these results are for reference only. More time and better cooling are needed to work around bugs and instabilities to find maximum stable clock. Results were achieved using Scythe Katana 3 cooler, 1.465V on the core, and HT (base clock) of 240 MHz. Value and Conclusion<table width="100%" cellpadding="5" cellspacing="0" id="result"> <tr><th>http://www.techpowerup.com/images/dollar.gif</th> <td>
</tr><tr> <th>http://www.techpowerup.com/images/thumbup.gif</th> <td>
<tr> <th>http://www.techpowerup.com/images/thumbdown.gif</th> <td>
<tr><th>8.9</th> <td>Looking back at the performance graphs, Athlon II X4 645 performed as expected in every way, both good and bad, and even throwing a few pleasant surprises along the way. Strong points for this processor are by far its reasonable price, and fast four physical CPU cores that in some cases perform almost unenviable good. For example in rendering benchmarks, where Athlon II X4 645 manages to match and even exceed the performance of Phenom II X4 945, without breaking a sweat. In its new C3 revision, Athlon II is seeing a nice drop in power consumption, in this case up to 20 W under load compared to the C2 revision of Athlon II X4 630, and that's not something you can ignore. With new optimizations Athlon II X4 645 reaches the same level as Intel's older Core 2 Quad models regarding power efficiency, but is still miles away from new Core i3 models. Luckily it has no problems matching their performance levels, except in single-threaded applications, so higher power consumption can be overlooked a bit. A Weak point for this processor, as well as the whole Athlon II X4 familiy is its unstable performance. Depending on the application, performance can vary from stellar to downright poor. This is without doubt caused by the lack of large L3 cache memory and inability to queue instructions, making the cores communicate with the much much slower system DRAM memory to get information on what to process next. Luckily, in the general computing world, where general purpose users dominate the population, and professional applications are very rare, this may not be such a handicap after all.<br /><br /> With a price of $ 122, fast clocked cores, strictly okay power consumption and backward compatibility to AM2+ sockets using DDR2 or DDR3 memory, it's really hard to find flaws in Athlon II X4 645. Even in its weakest segment, gaming performance, Athlon II X4 645 performed more than acceptable making up the lack of L3 cache with higher clocked cores. But I can't shake the feeling that AMD almost forced this fastest Athlon II X4 model to where it landed, and with just a little bit of luck, or should we say help from Intel. Although Intel can come up with an answer to Athlon II X4 645 quite easily, it chooses not to play AMD's aggressive pricing game and in doing so, it's making more than enough room on the market that AMD can exploit with processors such is Athlon II X4 645.</td></tr> <tr><th></th><td>http://www.techpowerup.com/images/recommended.gif</td></tr> </table> |
| All times are GMT. The time now is 11:56 PM. |
Powered by vBulletin® Version 3.8.6
Copyright ©2000 - 2013, Jelsoft Enterprises Ltd.