General Benchmark Databases
Owned Hardware
- (3 years)
- WD EALX 1.0TB harddisk, buy: 2011-03-18
- Toshiba 750G USB3.0 harddisk, buy: 2011-08-30
- LG W2353V 23″ LCD Display, buy: 2010-01-21
- GTR R450 450W PSU, buy: 2010-03-19
- WD WD3200AAKS 320G harddisk, buy: 2007-08-29, dead: 2012-03
- WD WD3202ABYS 320G harddisk, buy: 2008-09-06
- ASRock A780GMH/128M, buy: 2009-06-29
- (5 years)
- Seagate ST9160821AS 160G USB2.0 harddisk, buy: 2007-09-12
- WD 640G 6401AALS, buy: 2008-11-30 to 2013-11-30
- WD 640G 6400AAKS, buy: 2008-08-24 to 2013-08-24
- (No Warranty)
- LG W1942T 19″ LCD Display, buy: 2009-02-22
Software Timing
- True Image 9.7 high compress backup 32G Win7 C:, different harddisk, 12 min
- Windows 7 Upgrade RC to RTM, AMD Phenom II 3.1G + 8G RAM, 1h
USB 3.0 USB Keys
- Sandisk Extreme CZ80 32G, 178M read, 118M write?, $275
- ADATA S102 Pro 32G, 100M read, 44-49M write, $168
- Transcend Jetflash 760 32G, 65M read, 36M write, $175
- Kingmax UD-09 32G, 13-23M write?, $169
- Starram SLC 32G, 121M read, 145M write, easy to die?
Memory
- Sandy Bridge Memory Scaling: Choosing the Best DDR3
DDR3-1600 CL9 vs DDR3-1333 CL9, 3.5% faster in 7-Zip for Sandy Bridge 1333 CPU - Parallel Processing, Part 2: RAM and HDD
E6850, 1333Mhz, Dual Channel 800 less than 5% difference, average less than 2-3% difference even in WinRAR
Memory Card
- 市售幾款 SD UHS-1 速度測試比較 (2011-05)
- Sandisk Mobile Ultra 64G SDXC 30MB/s? (17MB/s write, FDBench 27MB/s)
- Sandisk Extreme HD 8G (20MB/s write)
- Tomshardware: SDHC min write speed
- Forum SDHC test result
- Olympus E300 SDHC card test
- Kingston 16G Class 10 MicroSDHC 14.4M write
- Kingston 16G Class 10 SDHC 11.4M write
- Japan Transcend SDHC benchmark
Harddisk
- Passmark Hard Drive Benchmarks
- Toms Hardware: Notebook Hard Drives: 750GB And 640GB Models Reviewed
- SSD Reliability
- Anandtech: Compare harddisk vs Intel SLC SSD at random 66% read, 33% write
- Intel SSDSA2SH032G1GN 32G SLC, DTR drop less than 5% upto 4 drives, drop 20% at 8 drives
- Seagate ST3300655SS 300G 15000RPM SAS, DTR drop 50% to 40%
- WD WD1000FYPS 1TB 7200RPM SATA (Green Power like 5400RPM), DTR droped 80%
- Tweakers.net Harddisk Benchmark – IOMeter Fileserver 64 I/O
- StorageReview.com’s Drive Performance Database (Testbed 4)
IOMeter File Server 64 I/O - StorageReview.com’s Drive Performance Database (Testbed 3)
File Server DriveMark 2002 = IOMeter test - Tom’s Hardware Interactive Harddisk Charts
- Tomshardware 3.5″ Max Read Transfer
- Tomshardware 2.5″ Max Read Transfer
- The Test Bed – Harddisk results
- Digit-Life: Detail benchmark on 2.5″ WD Scorpio ML80 WD1600BEVS and other harddisks
- PDF report compare CF/SSD/harddisk random write speed
CF/USB key are 2x to 10x slower than harddisk in small buffer (512 bytes) random write - Benchmark of Intel’s X-25M G2 SSD vs harddisks
- Info
- In March 2011, Western Digital announced it would acquire HGST, for $4.3 billion. The transaction is not expected to close until March 2012, due to the need to comply with remedy conditions set by the European Union.
- December 2011 – Seagate acquired Samsung’s HDD business.
Compress Software
- Hardware: Phoenom II 550BE 3.1Ghz x 2, 8G DDR2-800 RAM, WD 1002FAEX 1TB SATA3
- When unlocked, similar as a Phenom II X4 955 3.2Ghz (3.5% slower), Passmark ~4000
- Software: Windows 7 Ultimate, WinRAR 4.01 64bit, 7-Zip 9.20
- Test File: 456.7M VMware disk file
- WinRAR
- Fastest: 110.2M, 0:15 (24.1%, 3.5x faster)
- Fast: 96.6M, 0:52 (21.2%, 100% time)
- Normal: 91.0M, 1:01 (19.9%, 17% slower)
- Good: 90.4M, 1:12 (19.8%, 38.5% slower)
- Best: 90.1M, 1:23 (19.7%, 59.6% slower)
- 7-Zip
- Fastest: 95.8M, 0:31 (21.0%, 42% faster)
- Fast: 91.8M, 0:44 (20.1%, 100% time)
- Normal: 77.5M, 1:42 (17.0%, 2.3x slower)
- Maximum: 72.8M, 2:18 (15.9%, 3.1x slower)
- Ultra: 71.3M, 2:15 (15.6%, 3.1x slower)
- Hardware: Athlon64 X2 3800+ 2.0Ghz x2, 2G DDR2-667 RAM, WD2000JB 200G ATA100
- Software: WinXP Pro, WinRAR 3.7b7, compress to same disk, Multithread
- WinRAR on VMware 2.00GB w2k3r2
- Fastest, 921,526kB, 5m 15s
- Fast, 841,903kB, 10m (9.1%, 80M smaller)
- Hardware: Celeron 1.2G, 1G DDR266 RAM, WD2000JB 200G ATA100
- Software: WinXP Pro, WinRAR 3.6b7, compress to same disk
- WinRAR on VMware 25.5M disk
- Fastest, 4s, 3.53M/13.84%
- Fast, 10s, 2.62M/10.27% (34.8% better)
- Normal, 14s, 2.58M/10.12% (1.5% better)
- WinRAR on VMware 418M disk
- Fastest, 136s/2:16, 167.5M/40.07% (base, 1x time)
- Fast, 396s/6:36, 149.2M/35.7% (12.4% better, 2.9x time)
- Normal, 509s/8:29, 147.3M/35.24% (1.3% better, 3.7x time)
- Fastest decomp, 36s
- Fast decomp, 33s
- Normal decomp, 31s
- WinRAR on VMware 2.00GB w2k3r2
Display Card
- GeForce GTX 650 Ti
(GTX650 Ti and GTX660 is price/performance in proportional)
1920×1080 highest quality usually can get 35fps to 45fps- The NVIDIA GeForce GTX 650 Ti Review, Feat. Gigabyte, Zotac, & EVGA
Skyrim with highres pack has 80% difference for 2G RAM, others 1G & 2G almost same performance - Tomshardware Benchmark Results: The Elder Scrolls V: Skyrim
GTX 650 Ti vs GTX 560 vs GTX 460 192bit, Skyrim: HD7770 ~= GTX460?, GTX 650 Ti only 5% better than HD7770?
- The NVIDIA GeForce GTX 650 Ti Review, Feat. Gigabyte, Zotac, & EVGA
- Anandtech: GTX 460 256bit vs GTX 460 192bit (256bit is 10% better)
- GTX 560SE vs GTX 460
- GeForce GTX 460
- Skyrim Performance Test
GTX 460 vs GTX 560 etc
- Skyrim Performance Test
- HD 4850
- HD4850 vs HD4830
HD4850 = HD4830 + 10%, HD4830 ~= HD4770 + 3%
- HD4850 vs HD4830
- GeForce GT 240
- Entry Grade Display Cards
- HTPC最佳夥伴!入門HD5450超值現身(HK$270 for Sapphire HD5450 512M DDR3, 2012-02-20)
HD decoding for old PC should be ok (e.g. P4 3.0Ghz)
- HTPC最佳夥伴!入門HD5450超值現身(HK$270 for Sapphire HD5450 512M DDR3, 2012-02-20)
- GDDR3 vs GDDR5 real world performance maybe just 8%
- Adobe Premiere Pro CS5 benchmark
Compare nVIDIA CUDA performance, no AMD CPU here - Video Cards for Adobe Premiere CS5 and CS5.5
Explain the MPE in Adobe Premiere and the effect of nVIDIA CUDA cores (10x+ performance gain, require 1G RAM VGA card. Intel CPU with SSE4.1+ is additional 3x faster than AMD.) - NotebookCheck: Compare display chip performance
GTX 650 Ti ~= 95-100% (Especially after 4xAA) GTX 560 ~= 107% GTX 460 = 100% Crysis: Warhead (2008) Medium 0xAA/0xAF GTX 260 = 41 HD 5770 = 34 HD 4850 = 33 HD 4770 = 30 9800 GT = 28 9600 GT = 27 GT 240 = 17 Crysis: Warhead (2008) Ultra [1920x1200 EntQuality 4xAA] Nvidia GeForce GTX 650 Ti ($1180-1230) = 24.5 GTX 460 SE = 21.64 (8xAA DX10) ATI HD 5770 = 18.7 Nvidia GeForce GTX 650 = 17.8 Nvidia GeForce GTX 550 Ti = 17.7 ATI X1900 = 9 (low = 47) Crysis (2007) High [1024x768 High 0xAA/0AF] GTX 460 = 141% GTX 260 = 100% (2x 6pin power) Mafia 2 (2010) High [1366x768 high 0xAA 16xAF] Nvidia GeForce GTX 650 Ti ($1150) = ~83 Nvidia GeForce GT 650M (~GTX650 $850) = 61 ATI HD 4850 = 45.7 (2nd hand $450) Mafia 2 (2010) Ultra [1920x1080 high 0xAA 16xAF] Nvidia GeForce GTX 460 1G = 61 GTX 460 1G ~= 60.5? (worth ~$725) GTX 460 SE 1G ~= 55 (worth ~$675) HD 5830 ~= 54 HD 5770 1G = 50 (worth ~$600) Nvidia GeForce GTX 650 Ti ($1150) = ~50 Nvidia GeForce GT 650M (~GTX650 $850) = 37 Nvidia GeForce GT 240 GDDR5 = 32 (2nd hand $299) ATI HD 4850 = 29.5 (2nd hand $450, new $700-800) Need for Speed Shift (2009) High [1366x768 high/on 4xAA TriAF] ATI HD 4850 = 47 (2nd hand $450) Nvidia GeForce GT 650M (~GTX650 $850) = 45 ATI HD 6650M = 33 (2nd hand HD6670 $420) ATI HD 4670 = 25.6 (2nd hand $199)
(Compare 3D Mark 05 Standard 1024x768 / 3dmark06 1024x768 / 3dmark vantage 1024x768) nVidia GeForce GTX 660 Ti = (3dmark06:22997) Nvidia GeForce GTX 470M = 18.8x (23057, 3dmark06:16509 22.22x) - similar to GTX 560 Ti Nvidia GeForce GTX 560M = 18.46x (22648, 3dmark06:15223 20.49x) Nvidia GeForce GTX 660M = 16.98x (20831.7, 3dmark06:15117 20.35x) ATI HD 7850M = (3dmark06:14111 18.99x) Nvidia GeForce GT 650 = 17.3x (21266, 3dmark06:13719 18.46x) Nvidia GeForce GT 640M LE = 14x? (3dmark06: 9992 13.45x) ATI HD 6850M = 13.8x (16944) ~1.1x of GTX 550 Ti ATI HD 5850 = 12.44x (15260) ATI HD 4850 = 12.55x (15400, 3dmark06:9842 13.25x) ATI HD 5650 = 10.2x (12518) ATI X1950 Pro = 8x (~9300-10529 /avg9900,3dmark06;~5226 7.03x) Intel GMA HD4000 = 7.64x (9374, 3dmark06: 5322 7.16x) ATI HD 5550 = 637% (3dmark06,1280x1024:4735) ATI X1900(GT) = 580% (7117) ATI HD 5450 = 508% (3dmark06:3776) ATI HD 4350 = 4.85x (5952, 3dmark06: 3141 4.23x) Nvidia GeForce 305M = 417% (3dmark06:3101) * Nvidia GT218 ION2 = 311% (3821) run with Atom D525 Nvidia GeForce 9300M = 267% (3270) ATI HD 4290 (890GX) ATI HD 4250 (880G) = 279% (3423) * Nvidia GT218 ION2 = 253% (3098) run with Atom N450 ATI HD 4200 (785G) = 249% (3051) Intel GMA HD = 234% (2872, 3dmark06:1530) Core i3/i5 build-in ATI HD 4225 = 196% (2400) ATI HD 3300 (790GX) ATI HD 3200 (780G) = 189% (2324) Intel GMA 4500MHD = 100% (1227, 3dmark06:743) run with C2D SU2300 1.2Ghz ATI X1270 = 80% (983) Intel GMA X3100 = 66% (805) run with C2D 1.4-2.2Ghz Intel GMA 950 = 33% (407) run with C2D T2300 1.66Ghz * Intel GMA X3150 = 24% (298) run with Atom N4xx ~1.6Ghz * Intel GMA 950 = 20% (249) run with Atom N270 1.6Ghz VIA Crome 9 = 19% (227) Intel GMA 500 = 11% (135)
- Tom’s Hardware VGA Charts
- Updated: AMD 785G: The Venerable 780G, Evolved
Compare HD3200 and HD4200, FPS almost same, HD playback 25-50% better - New chipset AMD 880G review
Compare 785G, 880G, 890GX with benchmark
880G/890GX More Improved video playback quality - HD4770 Power Usage
HD4770: Core i7@3.2G, Load: 147/198w, Temp: 65
HD4830: Core i7@3.2G, Load: 148/213w, Temp: 90 - X1950 Pro vs HD4850 mini review
- ASUS X1950 Pro Power Usage
X1950 Pro: E6600@3.0G, Load: 156/220w - Radeon HD4770, 40nm
- 2009-04-30: HK$800
- 100-110% of 4830, power 91% (HK$900)
- 105% of 4850, power 83% (HK$1100)
- Sometimes > 4850
- ATI Radeon HD 4770: 40nm Goes Mainstream
- Geforce GTX 550 Ti vs GTX 560 Ti spec compare
- Anandtech: Radeon HD6790 benchmark (Include GTX 550 / GTX560)
- Tomshardware: AMD 790GX (HD3300) benchmark
30% better than 780G (HD3200) w/sideport memory, 60% better than 780G w/o sideport memory. HDD also seems around 2.5% better and power around 5% less.
LCD Monitor
- televisioninfo.com
Professional compare LCD/LED TV on color accuracy and other factors - Face to Face: Screens
A very professional website that use colormeter to measure LCD quality differences - X-bit labs: 20″ LCD Monitors Roundup. Part IV
- AnandTech – The LCD Thread
- 小熊在線—顯示器頻道
- OpenLGTV project
Customized LG TV firmware (Linux based)