- Status
[citation][nom]lashton[/nom]correct me if im wrong but i thought no intel CPU had OOO execution[/citation]
The Atom processors don't. All the rest have been doing it for years. Instruction reordering is expensive, and part of why modern x86 cores are relatively large, but it's also part of the reason that i3's huge IPC lead over the low-energy crowd.
AMD's low-power Bobcat chip, due Soon(TM) is a 40nm OOO part, which will also be competing with the Nano and Atom in the low-end notebook space.
Unfortunately, by the time the Nano DC launches, it won't be able to claim the fastest IGP anymore. AMD will have Bobcat out, which has a HD 54xx class GPU, and Sandy Bridge will be out too (the ones with faster IGPs are about on par with HD 5450). And of course, Llano is going to eat every IGP ever built for breakfast.
The Atom processors don't. All the rest have been doing it for years. Instruction reordering is expensive, and part of why modern x86 cores are relatively large, but it's also part of the reason that i3's huge IPC lead over the low-energy crowd.
AMD's low-power Bobcat chip, due Soon(TM) is a 40nm OOO part, which will also be competing with the Nano and Atom in the low-end notebook space.
Unfortunately, by the time the Nano DC launches, it won't be able to claim the fastest IGP anymore. AMD will have Bobcat out, which has a HD 54xx class GPU, and Sandy Bridge will be out too (the ones with faster IGPs are about on par with HD 5450). And of course, Llano is going to eat every IGP ever built for breakfast.
lpedraja2002
Distinguished
- Dec 8, 2007
- 620
- 0
- 18,990
Where's the street Fighter IV Benchie? I swear that's what I'm always looking for in the reviews but they took it out. They should change their game benchmarks for these lower-end performance IGPs.
They should be benchmarking older games like:
Valve games - CSS, HL2
COD2, CODMW
Games with UE3 which are great performers.
They should be benchmarking older games like:
Valve games - CSS, HL2
COD2, CODMW
Games with UE3 which are great performers.
megamanx00
Splendid
- Sep 3, 2008
- 8,216
- 2
- 31,465
Fusion will smack it and the VN1000 will call Fusion Daddy. I was kinda hopeful that VIA would provide something interesting, but perhaps with some polish and that 40nm CPU this will be a good setup for HTPCs, or not.
I did my calculations and would estimate the Via system at full cpu to be closer to 45W, compared to your 55W!
We'll see in the final version.
We'll see in the final version.
tom thumb
Distinguished
- Apr 18, 2010
- 181
- 0
- 18,690
looking at the benches none of these solutions are realistic to game on. If you have to tune down all settings to minimum (including resolution), and you still only manage 20fps... you just shouldn't be gaming. Games just plain loose their appeal. I think this was fairly obvious.
I am looking forward to VIA being able to do something in this one area. I have always felt that Atom was nearly useless, but I recognize the importance of the low voltage platform. All VIA needs is to get to market soon, be much better than this preview, and get on production hardware. All intel has to do to render VIAs efforts useless is to modify the I3 - say one core with HT. Something both companies must do is make them CHEAP.
Then there is AMD.
Then there is AMD.
[citation][nom]dman3k[/nom]NVidia needs to buy VIA while they are down so they can enter the CPU market and compete.[/citation]
+1 best idea i have heard yet.
idk what toms was thinking other then they seriously wanted to slap via in the face. throwing DX10 games at POS p/c's well since toms wants to skew things so bad like that then from now on i expect every bench mark to be max'd and only max'd no more of this 1080p b.s. screen size can be no smaller then 55" with 2 or 3 55" screens and the only game benchmark should be crisis or crisis2. no file size compression under 4.7gb either, no more itunes encoding full flac only and nothing that doesn't run full ddr3 or less then 16X sli and all the benchmarks have to be run and done at the same time for every application being tested.
seriously TOM's i don't like rigged comparrisons and will show you truely crippling results as watching a move burning a dvd while compressing the next one, while watching another dvd and playing a crisis and listening to flac while downloading all at the same time is how i roll every day. if you can't or won't do this i don't know how much longer any of us are going to continue bothering with this site's benchmarks.
anybody can skew things to show biased favor, especially me, i like to break every one no exceptions.
you haven't succeeded until you found the breaking point.
go all out or go home.
+1 best idea i have heard yet.
idk what toms was thinking other then they seriously wanted to slap via in the face. throwing DX10 games at POS p/c's well since toms wants to skew things so bad like that then from now on i expect every bench mark to be max'd and only max'd no more of this 1080p b.s. screen size can be no smaller then 55" with 2 or 3 55" screens and the only game benchmark should be crisis or crisis2. no file size compression under 4.7gb either, no more itunes encoding full flac only and nothing that doesn't run full ddr3 or less then 16X sli and all the benchmarks have to be run and done at the same time for every application being tested.
seriously TOM's i don't like rigged comparrisons and will show you truely crippling results as watching a move burning a dvd while compressing the next one, while watching another dvd and playing a crisis and listening to flac while downloading all at the same time is how i roll every day. if you can't or won't do this i don't know how much longer any of us are going to continue bothering with this site's benchmarks.
anybody can skew things to show biased favor, especially me, i like to break every one no exceptions.
you haven't succeeded until you found the breaking point.
go all out or go home.
- Dec 31, 2007
- 48,594
- 91
- 66,190
[citation][nom]f-14[/nom]idk what toms was thinking other then they seriously wanted to slap via in the face. throwing DX10 games at POS p/c's well since toms wants to skew things so bad...(more useless blather)...[/citation]Rigged? VIA said games. VIA said DX10.1. VIA said "Mainstream PC Platform". Tom's then proceeded to use the lowest-possible DX10 settings from its mainstream motherboard benchmarks in an effort to find some DX10 games that the IGP could pass.
From then your comments go off on a complete tangent so you should let us know: Are you really trying to say that Tom's hates VIA in your first sentence and then proceeding to say Tom's shills for VIA in your next paragraph?
From then your comments go off on a complete tangent so you should let us know: Are you really trying to say that Tom's hates VIA in your first sentence and then proceeding to say Tom's shills for VIA in your next paragraph?
iam2thecrowe
Glorious
- Jul 4, 2010
- 10,591
- 21
- 44,965
yes, if nvidia and via combined forces, they would be able to produce quite brutal cpu/gpu combo's. Between them and AMD, intel would then start to lose a fair bit of market share. Without that happening, Like always, VIA will fall short and offer a low end product that few people will want. Also VIA's graphics drivers/cards have always produced glitchy graphics and poor-average performance and I dont think more refined drivers have ever fixed their problems.
sonofliberty08
Distinguished
- Mar 17, 2009
- 658
- 0
- 18,980
i don't think VIA will like to sell them self to anyone , they try so hard to get everything they can on the desktop , so far they got the sound chip , network chip , graphic chip and now CPU ......
- Jul 26, 2008
- 3,873
- 520
- 23,840
Did they try changing the CPU-ID and rerunning some of those benchmarks? As an avid Via user (for home router / low end home servers) I've noticed that damn near all software on the market checks the CPU-ID to determine the code path. Most software is only optimized for Intel CPU's and sometimes AMD depending, most have no clue what a Via is and instead use a generic non-optimized code path. This will cause all sorts of weird performance issues, mostly because SSE type instructions won't be executed and instead everything will be done using plain x86 instructions.
On another note, WHY wasn't AES encryption tested in all this? One of the features of every Via CPU is onboard AES encryption, its called padlock. Because the encryption can be done with a handful of single instructions vs using a software solution its usually 10 ~ 100 times faster (I'm not joking about this). I've achieved real-time R/RW speeds on full disk encryption with a 1.8Ghz Via C7 CPU before, I imagine the Nano would be even faster.
On another note, WHY wasn't AES encryption tested in all this? One of the features of every Via CPU is onboard AES encryption, its called padlock. Because the encryption can be done with a handful of single instructions vs using a software solution its usually 10 ~ 100 times faster (I'm not joking about this). I've achieved real-time R/RW speeds on full disk encryption with a 1.8Ghz Via C7 CPU before, I imagine the Nano would be even faster.
- Jul 26, 2008
- 3,873
- 520
- 23,840
To further explain my above post. I use Linux on my Via systems and I found that I have to rebuild the Kernel of any distro I put on to get any enhanced features out of the system. Most installed software works ok but I find that rebuilding OpenSSL / OpenVPN / Apache and a few others nets me pretty big improvements once I recompile with it recognizing that the Via CPU has advanced x86 instructions. Problem with only have two big manufacturer's is all the software produced is only optimized for them and coders tend to get lazy and use CPU-ID to set a code path instead of feature set flags.
Nintendork
Distinguished
- Dec 22, 2008
- 464
- 0
- 18,780
sonofliberty08
Distinguished
- Mar 17, 2009
- 658
- 0
- 18,980
WarraWarra
Distinguished
- Aug 19, 2007
- 252
- 0
- 18,790
WOW it can play Crysis WOW.
Via might be waiting for some company that is going to use the Intel 1% of 22nm fab process before releasing ?
Via might be waiting for some company that is going to use the Intel 1% of 22nm fab process before releasing ?
Loock at this: VIA Nano L2200 1.6GHz (Single-Core Processor) + S3 Graphics Chrome 435 ULP (VIA Trinity Platform) play Crysis SP Demo
http://www.youtube.com/watch?v=mehGJsiBNA8
Crysis Demo Benchmark DX9 1280x720 all LOW details
VIA VB8003 (VIA Trinity) : 25,695 fps /Min FPS: 16.73 at frame 1947, Max FPS: 41.53 at frame 71/
http://extrahardware.cnews.cz/forum/viewtopic.php?f=38&t=10094
Total playble not a slide show
)
Randomacts
Distinguished
- Aug 5, 2010
- 795
- 0
- 19,010
- Dec 31, 2007
- 48,594
- 91
- 66,190
[citation][nom]tralalak[/nom]Crashman: Please write in youtube: VIA Nano + S3G Chrome 435 ULP play Crysis and Xpand Rally.[/citation]I didn't try DX9 mode because DX10.1 is supposed to be one of its major features. Then again, a min FPS of 16.73 in DX9 mode could be 1/2 second of 0FPS (1 frame) and 1/2 second of 32FPS (16 frames in 1/2 second), which is still unplayable. That's why the site generally sticks to a 20FPS lowest minimum or a 40FPS lowest average as its frame of reference for playability.
- Status
TRENDING THREADS
-
-
Question New installed W11 on sdd, D drive (HDD) goes to the 100% in seconds
- Started by Emily_Vin
- Replies: 8
-
-
Facebook
Twitter
Instagram