打印

[硬件] 10W低端赛扬已可媲美65W高端酷睿2

10W低端赛扬已可媲美65W高端酷睿2

This fast?!

这两年无论是AMD还是Intel,处理器性能的提升幅度都小得可怜,但如果把眼光放长远一些,历史的进步还是相当猛的,可能会超出很多人的想象。
比如一边是今年的Celeron J1850,Bay Trail-D入门级桌面平台,22nm工艺,四核心,主频2.0GHz,热设计功耗10W;另一边是2008年初的Core 2 Duo E8200,扣肉时代的高端桌面产品,45nm工艺,双核心,主频2.66GHz,热设计功耗65W。
将它们俩放在一起,孰强孰弱呢?
作为基准性能测试软件开发商,PassMark Software的官方网站上还有一系列的处理器、显卡、硬盘性能数据库。通过查询可知,Celeron J1850的平均得分为1814,Core 2 Duo E8200则是平均1845。
这意味着什么?五年多过去了,你现在只需要当初不到六分之一的功耗,就可以得到几乎同样的性能,如今的入门级产品已经媲美当年的高端型号了!
此外,Ivy Bridge家族的双核心Celeron G530T 2.0GHz/35W平均得分为1604,换了“Atom芯”的新赛扬都要比它强13%呢。
这一系列还有更高端的Pentium J2850/J2900 2.41GHz,得分超过2000应该不成问题。


时代的进步



目前性能、性价比最好的一些处理器



http://news.mydrivers.com/1/284/284696.htm

TOP

呢邊10W 個邊就200W

TOP

引用:
原帖由 ccw 於 2013-12-1 11:27 發表
This fast?!

这两年无论是AMD还是Intel,处理器性能的提升幅度都小得可怜,但如果把眼光放长远一些,历史的进步还是相当猛的,可能会超出很多人的想象。
比如一边是今年的Celeron J1850,Bay Trail-D入门级桌 ...
AMD A6-5200 APU (25W TDP, 4C, 2.0GHz) 2,446
AMD A4-5000 APU (15W TDP, 4C, 1.5GHz) 1,912
AMD A6-1450 APU (8W TDP, 4C, 1.4/1.0GHz) 1,644
AMD E1-2500 APU (15W TDP, 2C, 1.4GHz) 837

不過就算係數字快過C2D, single-thread performance (低階CPU比較重要的metric) 都係差1倍,
PassMark基本上係perfect scaling.

TOP


用佢做ser咪快過我部g530

TOP

引用:
原帖由 cheungmanhoi 於 2013-12-1 18:43 發表

用佢做ser咪快過我部g530
should be

TOP

引用:
原帖由 qcmadness 於 2013-12-1 04:55 發表

AMD A6-5200 APU (25W TDP, 4C, 2.0GHz) 2,446
AMD A4-5000 APU (15W TDP, 4C, 1.5GHz) 1,912
AMD A6-1450 APU (8W TDP, 4C, 1.4/1.0GHz) 1,644
AMD E1-2500 APU (15W TDP, 2C, 1.4GHz) 837

不過就算係數字快 ...
Not exactly, if the quad core score is counted by 1.99GHz,
then with single core + turbo boost, the difference should be narrower.
And for J2900, 10W TDP but 2.4GHz quad core, it should be much faster.


[ 本帖最後由 ccw 於 2013-12-1 11:05 編輯 ]

TOP

引用:
原帖由 cheungmanhoi 於 2013-12-1 06:43 發表

用佢做ser咪快過我部g530
Not really for J1850, as they were taking G530T in the test,
G530 itself can reach 22XX in passmark, but still it maybe overtaken by J2900 later....

TOP

引用:
原帖由 ccw 於 2013-12-1 23:04 發表

Not exactly, if the quad core score is counted by 1.99GHz,
then with single core + turbo boost, the difference should be narrower.
And for J2900, 10W TDP but 2.4GHz quad core, it should be much faste ...
For similar mark with dual-core E8400, the single-thread performance difference is 100%, turbo boost by 20% will narrow the difference to about 67% times faster (2/1.2 = 1.67).

You mean J2900 vs A6-5200?
AMD deliberately disables turbo-core for high-end Jaguar cores, so the real difference is not that large.

TOP

And for Pentium J2900, the expected mark is about 2218

TOP

引用:
原帖由 qcmadness 於 2013-12-1 11:18 發表
And for Pentium J2900, the expected mark is about 2218
Recalling the Celeron G530 I got in early 2012, it took like 22XX marks in passmark,
and is now overcome in the multithread score by this 10W TDP monster.........time to build the new router/nas/server with ATOM

TOP

引用:
原帖由 ccw 於 2013-12-2 02:45 發表

Recalling the Celeron G530 I got in early 2012, it took like 22XX marks in passmark,
and is now overcome in the multithread score by this 10W TDP monster.........time to build the new router/nas/serv ...
In fact, a dual-core Bay-Trail or dual-core Jaguar is more than enough for router / nas / low-end server.

TOP

引用:
原帖由 ccw 於 1/12/2013 11:27 發表
This fast?!

这两年无论是AMD还是Intel,处理器性能的提升幅度都小得可怜,但如果把眼光放长远一些,历史的进步还是相当猛的,可能会超出很多人的想象。
比如一边是今年的Celeron J1850,Bay Trail-D入门级桌 ...
want to replace it for my main computer, turn it to 24 hrs on machine
(using G840 now)

TOP

引用:
原帖由 bebird 於 2013-12-1 23:06 發表

want to replace it for my main computer, turn it to 24 hrs on machine
(using G840 now)
I doubt that it would suffice daily use at this time.
You'd better stick with G840 for a better performance, assembling yet another Bay Trail server would be more appealing.

TOP

睇下佢d版要幾多錢 可以諗下換換佢

TOP