卡長--由檔板開始測量，全長總共28公分!



SAMSUMG
K4G10325FE-HC04
由於是NVIDIA第一款高階世代採用GDDR5的產品(之前RE過又RE的不計算在內)，可以看到這款GTX480用的是-0.4ns的SAMSUMG GDDR5顆粒，只有正面才有記憶體顆粒，全部為128MB乘上12顆組成1536MB的顯示記憶體規格，記憶體匯流排為單顆32bit乘上12顆等於官方標定的總384bit匯流排規格，目前可以在網路上查到的此款0.4nsGDDR5運算頻率為5Ghz，而官方顯示記憶體實際時脈，GDDR5運作時脈為1848MHz，也就是實際在乘上2倍等於3696Mhz，這意味著!!顯示記憶體還能夠超很大@Q@?



裸卡一覽



GPU與記憶體的供電模組







此次的研發核心代號-GF100 375 A3版本
製程完成時間為2010年的第四周



記憶體導熱模組-背面



記憶體導熱模組-正面







8+6Pin模組保護罩?



4Pin風扇插頭



導熱器背面-5根熱導管



側面



導管下拍



導管上拍



導熱器外露下前方

風罩-很薄，玩家拆卸卡榫時要注意力道，不然卡榫容易斷裂。



外露導熱模組與風罩示意照







風罩拆卸下內部全貌



這次測試的SLI平台





單卡待機耗電量219~231W之間跳動，抓最高耗電量拍下此張照片



單卡最高耗電功率524~564之間跳動(風扇轉速85%時拍照)



SLI模式測得待機耗電量255~278W之間跳動，一樣抓取最高耗電拍照



SLI模式最高測得耗電功率588~624W之間跳動



由此結果換算實際耗電量--
GTX 480待機耗電量-278W-231W=47*0.8實際轉換率為20多W~37.6W之間(單卡)
GTX 480滿載耗電量-624W-278W=346*0.8實際轉換率為250多W~276W之間 (SLI模式下)
GTX 480滿載耗電量-564W-231W=333*0.8實際轉換率一樣為250多W~266.4W之間(單卡模式下)

這跟官方標的250W數字是接近的，也有可能是我周邊平台的某些元件有所關係
另外，此次測試的POWER跟上次測試GTX 295是一樣的銀欣850W，所以大家不要在猜了，什麼買這次費米要使用超過1200W的POWER!!?根本是無稽之談XD
因為SLI模式下，我放著跑FurMark 6小時都沒事，睡起來關掉FurMark繼續測試其他軟體都還是嚇嚇叫XD


--------------------------------------------------------------------------------------------
溫度測試--
測試軟體FurMark
此時室溫23度，無冷氣~
直接上雙卡SLI的待機溫度比較客觀，最熱也不過如此了(笑)
測得條件-開機40分鐘後，於桌面下都不作動任何程式!




單卡模式下，FurMark測試10分鐘，最終穩定於92度c
GTX 480 32億的電晶體對比ATI 5870似乎沒有比較熱XD



SLI模式下測試看看~
一樣跑FurMark測試10分鐘，最終定於91度c，跟上面結果誤差正負不到一度!


--------------------------------------------------------------------------------------------


遊戲與測試軟體實測~

3DMark Vantage

GTX 480單卡X模式下分數9567





在來直接上SLI模式下的所有Vantage分數
一般模式下[P]
得分34646




中階模式[H]
得分25238



最高階模式[X]
得分17655


那如果在最高階模式[X]下又開啟16X AF呢?
得分17686 @@..測到這邊，我一直認為分數一定是錯誤了，怎麼會開啟16XAF後分數不降反升??
連續的Try..Try..Try到我媽都快不認識我了
但是得到的結果竟然都是一樣的~"~..
意思是費米的硬體架構就是希望給大家開免錢的16XAF?
我不太相信..所以下面還會驗證一下~

DX11測試軟體-Heaven Benchmark v2.0
為了驗證16XAF是否開免錢，可以開到連我媽都嚇到，特地用Heaven Benchmark 1.0來驗證(因為之前網路有人質疑2.0是為了NV所開發)
廢話不多說，直接來吧~

GTX 480 SLI模式下
第一個模式解析度為1680*1050得分為105.5



第二張圖一樣解析度為1680*1050，唯一不同的是開啟16XAF模式得分...105.4
效能折損幾乎可以用誤差值來計算~"~


我不信邪...
繼續測試開啟24吋螢幕的最高解析度來跑
1920*1200得分89.2



開始16XAF之後..得分88.6
GTX 480"好像"有喘一點點點點點了XD


但是真的得到驗證，此張卡買回去後，在一般狀況下，就把16XAF直接開下去吧..反正不用錢! (狂笑~)


------------------------------------------------------------------------------------------------------

接下來換Crysis上陣吧~

一開始一樣SLI模式，但是由於驅動程式的關係無法在64位元下面作測試(第一版驅動嘛)，不然我想在64位元下分數會更高!!

Run #1- DX10 1680x1050 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 68.92
Run #2- DX10 1680x1050 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 64.27
Run #3- DX10 1680x1050 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 63.395
Run #4- DX10 1680x1050 AA=16xQ, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 59.095
Run #5- DX10 1900x1200 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 64.475
Run #6- DX10 1900x1200 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 55.61
Run #7- DX10 1900x1200 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 54.655
Run #8- DX10 1900x1200 AA=16xQ, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 48.545


這張我是連最高解析度模式都開啟16XQ模式，一般玩家不用這麼變態V.V

Run #1- DX10 2560x1600 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 33.05
Run #2- DX10 2560x1600 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 32.74
Run #3- DX10 2560x1600 AA=16xQ, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 9.125



CRYSIS 16XAF開啟

既然上面也測試過16XAF了，Crysis這款當然也不用給他客氣了，直接開啟..我管他老媽要嫁給誰!
成績請玩家自己解讀吧，分數變不變態，不是我說了算XD

Run #1- DX10 1680x1050 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 67.44
Run #2- DX10 1680x1050 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 60.855
Run #3- DX10 1680x1050 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 59.905
Run #4- DX10 1900x1200 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 60.825
Run #5- DX10 1900x1200 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 51.08
Run #6- DX10 1900x1200 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 50.16


Run #1- DX10 2560x1600 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 30.635
Run #2- DX10 2560x1600 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 30.04

然後是2560*1600這種最高解析度的8XQ模式也開下去...接近一秒30多張，現在怕的是你用的螢幕不夠大，不夠High而已了XD

Run #1- DX10 2560x1600 AA=8xQ, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 29.485


再來我感覺玩的不太爽，都開了16XAF效能還是不降，索性在開啟8X超級取樣，這下費米也應該要倒了吧XD



開啟反鋸齒8X取樣技術，費米終於快投降了XD
但是看到的結果，除了16XQ模式下不能順暢的進行遊戲，其他都還有30多張=..=
這時候我想到一句話:不要問，很恐怖一"一

Run #1- DX10 1680x1050 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 67.44
Run #2- DX10 1680x1050 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 31.98
Run #3- DX10 1680x1050 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 31.775
Run #4- DX10 1680x1050 AA=16xQ, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 17.6
Run #5- DX10 1900x1200 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 60.9
Run #6- DX10 1900x1200 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 25.075
Run #7- DX10 1900x1200 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 24.95
Run #8- DX10 1900x1200 AA=16xQ, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 13.955
Run #9- DX10 2560x1600 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 37.695
Run #10- DX10 2560x1600 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 13.995
Run #11- DX10 2560x1600 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 14.48

----------------------------------------------------------------------------------------------------------------

後面這邊開始測試單卡成績~索性的8X取樣也不關了，直接測吧~"~
分數如下:


Run #1- DX10 1680x1050 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 45.44
Run #2- DX10 1680x1050 AA=4x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 38.575
Run #3- DX10 1680x1050 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 37.66
Run #4- DX10 1680x1050 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 37.03
Run #5- DX10 1680x1050 AA=16xQ, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 33.75
Run #6- DX10 1900x1200 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 36.875
Run #7- DX10 1900x1200 AA=4x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 31.16
Run #8- DX10 1900x1200 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 30.47
Run #9- DX10 1900x1200 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 29.995
Run #10- DX10 1900x1200 AA=16xQ, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 27.245


Run #1- DX10 2560x1600 AA=No AA, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 22.015
Run #2- DX10 2560x1600 AA=8x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 18.155
Run #3- DX10 2560x1600 AA=16x, 32 bit test, Quality: VeryHigh ~~ Overall Average FPS: 17.9

------------------------------------------------------------------------------------------------------------

再換Crysis WARHEAD上陣吧~

單卡成績平均都40多張以上，而且也都開啟16XAF，不關了

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 0xx
==> Framerate [ Min: 46.20 Max: 93.13 Avg: 74.82 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 4xx
==> Framerate [ Min: 46.49 Max: 77.15 Avg: 61.54 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 8xx
==> Framerate [ Min: 45.24 Max: 76.22 Avg: 60.33 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 16xx
==> Framerate [ Min: 44.33 Max: 74.96 Avg: 59.43 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 16xQx
==> Framerate [ Min: 37.07 Max: 67.02 Avg: 52.18 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 0xx
==> Framerate [ Min: 44.92 Max: 75.13 Avg: 62.06 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 4xx
==> Framerate [ Min: 36.78 Max: 62.82 Avg: 50.31 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 8xx
==> Framerate [ Min: 36.49 Max: 62.61 Avg: 49.29 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 16xx
==> Framerate [ Min: 35.63 Max: 61.85 Avg: 48.78 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 16xQx
==> Framerate [ Min: 29.23 Max: 55.21 Avg: 42.59 ]

------------------------------------------------------------------------


SLI模式

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 0xx
==> Framerate [ Min: 43.98 Max: 117.26 Avg: 85.12 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 8xx
==> Framerate [ Min: 46.41 Max: 111.98 Avg: 84.80 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 16xx
==> Framerate [ Min: 47.03 Max: 110.91 Avg: 84.72 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 16xQx
==> Framerate [ Min: 46.66 Max: 99.56 Avg: 80.84 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 0xx
==> Framerate [ Min: 46.94 Max: 113.56 Avg: 85.36 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 8xx
==> Framerate [ Min: 46.77 Max: 98.79 Avg: 80.48 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 16xx
==> Framerate [ Min: 46.60 Max: 98.03 Avg: 79.78 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 16xQx
==> Framerate [ Min: 46.88 Max: 86.37 Avg: 70.14 ]


GTX 480 SLI模式開啟16XAF，又是一次在高解析度下不降反升的狀態V.V

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 0xx
==> Framerate [ Min: 35.80 Max: 114.07 Avg: 85.18 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 4xx
==> Framerate [ Min: 46.25 Max: 107.84 Avg: 83.87 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 8xx
==> Framerate [ Min: 46.62 Max: 105.04 Avg: 82.70 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1680 x 1050 AA 16xx
==> Framerate [ Min: 46.94 Max: 104.92 Avg: 82.79 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 0xx
==> Framerate [ Min: 46.70 Max: 108.11 Avg: 83.80 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 4xx
==> Framerate [ Min: 46.47 Max: 96.79 Avg: 78.45 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 8xx
==> Framerate [ Min: 47.45 Max: 93.09 Avg: 76.06 ]

DirectX 10 GAMER 3X @ Map: ambush @ 0 1920 x 1200 AA 16xx
==> Framerate [ Min: 45.43 Max: 91.70 Avg: 75.10 ]

----------------------------------------------------------------------------------

或許又有人會質疑，這面這些都是NV綁標的Game?
不盡公平?
好吧~

那我就找DX11的源頭開發商，這樣總不會再有人說話了吧?
微軟的SDK DX11開發套件~
以下這幾張開放給5870 CF或是5970挑戰看看，當然..你們可以使用最新的10.3版驅動程式!!
以下我特別不用超高解析度來作測試，一般20吋螢幕就能辦到了~
特別再提示一下，這幾款的畫面，有幾個是超越5倍的(不是30%這種)畫面張數 0.0+
所以真的請小心應戰..

DX10.1 兩款





DX11 4款











----------------------------------------------------------------------------------------------

結論:
想用大尺寸螢幕玩Game的玩家，你們可以準備升級了~
用2560*1600的解析度玩遊戲才是王道XD

辛苦了~~..................

我做一兩個測試就煩了
真配服z大

忘了說
這些會不會太清淡了些?

Z大把AF跟AA搞混了 ...
剛剛還在想AF效能不是沒耗那麼多

推,謝謝z大

不知道還有沒有其他遊戲測試呢?

這次的旗艦GTX480供電系統出乎意料的普通,反而有點不太像卡王XD

不過我還是會買啦,不知道何時買得到?

感謝分享

國外的網站也開始貼測試結果出來

小弟得到的結論是 (以玩游戲為出發點, 不是軟體開發者)

GTX480 在高解析度 (1920x1200/2048x1536) 比5870好一些
1. 價格高一點
2. 耗電高很多
3. 噪音/熱大很多
4. 如果去年9月上市, 絕對給90分, 比5870晚半年出來只能給70分
5. PSU至少要500W足瓦, 如果想超CPU還要更大

GTX480 SLI 比 5970 快很多
1. 價格只有最高階玩家吃得下 (USD$1000)
2. 耗電高得嚇人
3. 噪音/熱也大得嚇人
4. PSU至少要750W足瓦, 如果想超CPU還要更大
5. 目前是高階玩家首選 (SLI 效率高, 單卡1.5G RAM對2048 x 1536有極大幫助)