[Digitimes] AMD updates product roadmap for 2014 and 2015

[krumme]

I don't see why they can't just use quad channel DDR3 like the Xbox One.

It seems like the absolute easiest solution.

They also need to fix their absolutely archaic and horrendously performing integrated memory controller (compared to Intel's)

The only reason Intel still uses dual channel DDR3 on their laptops and desktops is because they have an integrated memory controller than performs about 2x as well as AMD's, making quad channel have no real world benefit for them in desktop applications.

Aparently its very difficult to get that good memory performance. As far as i remember Intel cache and memory management have been far superior. How is that? - is it more production than design related?

 

[Insert_Nickname]

VIA Nano 2 could have compited with Brazos, the problem was VIA started with single core 65nm process and when 40nm dual core was avalible it was too late. Combine that with crappy drivers/crappy igp and you got a recipe for failure.
VIA should have looked for Nvidia chipsets after the Intel C2D/Atom+Nvidia partnership was over.

Except the Nano2 purposefully wasn't targeted at the consumer market, but the embedded market. I have seen one consumer focused board using a VIA CPU since the socket 7 days (15+ years ago). No laptops or netbooks what-so-ever. The point is that VIA wasn't/isn't trying to compete with Intel/AMD, but carve out a little niche for themselves.

But you're quite right about their really crappy IGPs. But Intel is just as guilty of making crap IGPs. Ever used Intel "Extreme" Graphics (1/2)?

 

[Insert_Nickname]

But you're quite right about their really crappy IGPs. Crapy was all VIA Chrome9 Series IGP with game performance but VIA Chrome9 HCM (VIA VX855 MSP all in one chipset) a VIA Chrome9 HD (VIA VX900 MSP all in one chipset) can HW accelerate MPEG-4 AVC H.264 / WMV9 / VC1 but Intel "Extreme" Graphics not. But also never achieve in consumer SOHO!

I don't think you understand. You're comparing a relatively modern IGP with a 10+ year old one. Of course the Chrome9 is going to outperform a 10+ year old IGP and have more features. VIA IGPs have however always suffered from poor driver support.

I was just using the very old "Extreme" Graphics to illustrate that even Intel sometimes makes poor products. The best that can be said for the "Extreme" Graphics was that it put an image on the monitor, but even that much was spotty sometimes.

 

[Insert_Nickname]

VIA IGPs have however always suffered from poor driver support. = it does not matter because all this IGPs dont have gaming experience.

Who said anything about gaming? It does matter when it affects stability. I haven't even mentioned their chipset/IDE/AGP drivers yet. About the worst drivers I have ever had the misfortune of running across have been VIA and SiS ones.

Weather or not VIA drivers have improved I simply don't know, because I retired the last VIA-chipset based PC in my family in 2010 (or rather it retired itself, power-surge fried PSU/Mainboard). It was an AthlonXP T-bred A 1800+ in a KT400/VT8235-based mainboard from 2002. Haven't used anything VIA since...

 

[krumme]

Who said anything about gaming? It does matter when it affects stability. I haven't even mentioned their chipset/IDE/AGP drivers yet. About the worst drivers I have ever had the misfortune of running across have been VIA and SiS ones.

Weather or not VIA drivers have improved I simply don't know, because I retired the last VIA-chipset based PC in my family in 2010 (or rather it retired itself, power-surge fried PSU/Mainboard). It was an AthlonXP T-bred A 1800+ in a KT400/VT8235-based mainboard from 2002. Haven't used anything VIA since...

You guys are to young and spoiled.
In the nienties you needed solid soldering skills and deep format c: competences to get a 3d card going after an installation of some random beta drivers ruined your win95b installation. The times before 3dfx.
But we were happy because we could boot into 64.000 colors - we knew how to edit our system files.
Then we remembered the 80 ties without mouse but autoexec.bat and config.sys just to reach wordperfect and write simple text and we felt spoiled and old.

(And in the nienties we had to draw 20 meters of punched tape across our ass just to see the first prompt)
Edit seventies that is

 

[Shivansps]

• Power Consumption (max TDP):
1. Intel Bay Trail - D (10W)
2. AMD Kabini (15 - 25W)
3. VIA QuadCore CN-R (premise 15 - 25W) + VX11H (5.8W)

Considering VIA is more embedded, and its on pico and nano itx im suprised they did nor remove the NB yet.

 

[mikk]

KV SPECTRE DESKTOP 100W (1305).?....

http://www.sisoftware.eu/rank2011d/...f9c4f5d3bb86b395edd0e1c7a2c7facaec9fa29a&l=en


http://www.sisoftware.eu/rank2011d/...e7d6e0d0e3d7e4c2b08dbd9bfe9ba696b0c3fec6&l=en

Number of Devices / Threads 1 / 2


That's a Dual graphics system most likely.

 

[Insert_Nickname]

You guys are to young and spoiled.

I'd just like to point out that I have been into computers/IT since 1990, earlier if you count it as my hobby.

Had my fair share of Win 3x/95 issues. Back then, just playing Doom was immense because after a 2 hour battle with IRQ conflicts and what not, you finally managed to get the sound working...

 

[krumme]

I'd just like to point out that I have been into computers/IT since 1990, earlier if you count it as my hobby.

Had my fair share of Win 3x/95 issues. Back then, just playing Doom was immense because after a 2 hour battle with IRQ conflicts and what not, you finally managed to get the sound working...

Lol. An entire generation learned trial and error without having the sligtest idea what they were doing. Sometimes we had to break it just to see what happened to get just a small step further understanding.

 

[Insert_Nickname]

Lol. An entire generation learned trial and error without having the sligtest idea what they were doing. Sometimes we had to break it just to see what happened to get just a small step further understanding.

I think the intricacies of the 8259 interrupt controller(s), config.sys, himem.sys and emm386.sys would properly leave younger enthusiasts screaming. I know, I did sometimes. Know what, we managed. Not equally successful every time, but most of it... LOL...

It was just an example after all...

 

[NTMBK]

VIA Nano (65nm) was better as Intel Atom (45nm).
VIA Nano X2/QuadCore (40nm TSMC) is (was) better as AMD Bobcat (40nm TSMC) or Intel Atom (32nm).

Oh Tralalak, the world's last remaining Via fanboy!

On what planet was Nano X2 better than Bobcat?

And the power consumption was far worse, too: http://www.tomshardware.com/reviews/nettop-performance-review,3211-7.html

The Nano X2 was in no way better than Bobcat. The CPU was quite a lot slower, and the iGPU was a disaster. It doesn't even support DX10.

 

[NTMBK]

[FONT=&quot]Nice experiment... but:[/FONT][FONT=&quot]

AMD E-450 1.65GHz "Bobcat" microarchitecture [/FONT][FONT=&quot](+450MHz than VIA => +37,50% CPU clock)[/FONT][FONT=&quot]
AMD E-350 1.6GHz "Bobcat" [/FONT][FONT=&quot](+400MHz than VIA => +33,33% CPU clock)[/FONT][FONT=&quot]
VIA Nano X2 U4025 1.2GHz "Isaiah"
[/FONT][FONT=&quot]source:[/FONT][FONT=&quot] http://www.tomshardware.com/reviews/nettop-performance-review,3211-2.html[/FONT][FONT=&quot]


The same CPU clock (1.6GHz):

Bulldozer vs. Merom vs. Bobcat: 2 cores, 1,6 GHz + NetBurst 1 core, 3,46 - 3,6 GHz, + VIA Nano X2 @1,6 GHz[/FONT]

• [FONT=&quot]OggDropXPd:[/FONT][FONT=&quot] VIA Isaiah have better performance than AMD Bobcat[/FONT]
• [FONT=&quot]x264 benchmark:[/FONT][FONT=&quot] VIA Isaiah have better performance than AMD Bobcat[/FONT]
• [FONT=&quot]Cinebench R11.5:[/FONT][FONT=&quot] VIA Isaiah have better performance than AMD Bobcat[/FONT]
• [FONT=&quot]POV-Ray 3.7 RC3:[/FONT][FONT=&quot] VIA Isaiah have better performance than AMD Bobcat[/FONT]
• [FONT=&quot]WinRAR 4.1 beta:[/FONT][FONT=&quot] VIA Isaiah have better performance than AMD Bobcat[/FONT]

[FONT=&quot]
[/FONT][FONT=&quot]source:[/FONT][FONT=&quot] http://diit.cz/clanek/bulldozer-vs-merom-vs-bobcat-2-jadra-16-ghz-netburst-1-jadro-346-36-ghz-via

[/FONT][FONT=&quot]GPU in the tomshardware test was only old DX9 VIA Chrome9 HD (80nm lithography[/FONT][FONT=&quot]) => power consumption: (VIA Nano X2 (40nm [/FONT][FONT=&quot][FONT=&quot]lithography[/FONT]) + VIA VX900H MSP [/FONT][FONT=&quot](80nm [/FONT][FONT=&quot][FONT=&quot]lithography[/FONT])[/FONT][FONT=&quot] vs. AMD E-350/450 [/FONT][FONT=&quot][FONT=&quot](40nm [/FONT][FONT=&quot][FONT=&quot]lithography[/FONT])[/FONT] + [/FONT][FONT=&quot]AMD A50M FCH [/FONT][FONT=&quot][FONT=&quot](65nm [/FONT][FONT=&quot][FONT=&quot]lithography[/FONT])

* VIA Chrome 645/640 iGPU DirectX 11 is [/FONT][/FONT][FONT=&quot](40nm [/FONT][FONT=&quot][FONT=&quot]lithography[/FONT]) in the all-in-ine chipset VIA VX11H MSP.
[/FONT]

Those figures are both true, and meaningless. If you clocked the VIA part higher then sure, maybe it would perform better than Bobcat- but the power consumption would become even worse, and Bobcat already had lower power consumption at 1.6GHz than Nano X2 had at 1.2GHz. Different uArches target different clocks- the narrower Bobcat core can clock higher at the same TDP than the Nano X2 can.

 

[NTMBK]

I.

1. TDP isnt power consumption
2. AMD, Intel and VIA have differently method to TDP
3. VIA Isaiah is much bigger as AMD Bobcat
3.1 128Kb (64 Kb + 64 Kb) L1 cache per core vs. 64Kb (32 Kb + 32 Kb) L1 cache per core
3.2 1024 Kb L2 cache per core vs. 512 Kb L2 cache per core
3.3 3-issue vs. 2-issue
3.4 SIMD up to SSE 4.1 vs. SIMD up to SSE 4A
3.5 VIA PadLock kryptografic coprocessor vs. nothing
4. VIA Nano L4350E 1.6+GHz (1.73GHz capable) max 27,5W TDP vs. AMD E-350 1.6GHz/E-450 1.65GHz max TDP 18W
5. VIA VX11H MSP max TDP 5,8W vs. AMD A50M (Hudson M1) max TDP 5,9W
6. (4 + 5) 27,5W + 5,8W = 33,3W TDP vs. 18W + 5,9W = 23,9W TDP


II.



Can be? But ok. So hypothetical argument TDP vs. TDP (Isaiah vs. Bobcat).
VIA Nano L4350E 1.6+GHz (1.73GHz capable) 27,5W TDP = cca. hypothetical AMD Bobcat E2 at 2100-2200MHz.

But

The same 27,5W TDP than VIA Nano L4350E 1.6+GHz (1.73GHz capable) have VIA QuadCore L4700E 1.2GHz (1.46GHz capable) 27,5W TDP.

VIA QuadCore L4700E 1.2GHz (Isaiah) vs. AMD E-350 1.6GHz (Bobcat)
source: http://diit.cz/clanek/test-via-quadcore-l4700e-1200-mhz/testy-renderovani-cinebench-pov-ray

CONCLUSION:

I.
REBUT

II.
REBUT

Best regards

Tralalak

P.S.

source: http://www.anandtech.com/show/4023/the-brazos-performance-preview-amd-e350-benchmarked/3

I didn't talk about TDP, I talked about measured power consumption. The Tom's Hardware article has a section on power consumption, I recommend that you read it. The 1.2GHz Nano X2 had the second worst power consumption of all (even higher than a Sandy Bridge Celeron at 1.2GHz), with only the Sandy Bridge i3 having higher power consumption.

And sure, keep quoting cache sizes and issue width. It's an interesting detail, but what really matters is performance within a given power envelope at a given price, and the Nano X2 loses to Bobcat.

 

[Insert_Nickname]

• H.264 AVC 1080p DXVA HW acceleration: max. 24W (90% of time is spent on 23W)

Who is that supposed to impress? I have a fully fledged desktop A10-6800K-based mediacentre that can do that at ~28W. Never breaks 30W. That's an impressive 4W more...

Not that I'm arguing, just saying...

 

[NTMBK]

I have of course ZOTAC ZBOX nano VD01 with VIA Nano X2 U4025 1.2GHz (max TDP 18W) at home. VIA Nano X2 U4025 1.2GHz is only for ZOTAC ZBOX nano VD01 Series. VIA Nano X2 E-Series L4300E 1.2+GHz (1.46GHz capable) have max TDP 13W.

I mean that this Tom's Hardware article in section on power consumption have a repugnant verdict.



Arcsoft's TotalMedia Theatre 5 and VLC dont supported VIA ChrometionHD 2.0 video engine. => that test was only with software decoding => that test was decoded only with hard power VIA Nano X2 1.2GHz 95% CPU utilization => Video Playback (H.264, 720p) 35,3W D:

VIA supported Windows Media Player 11, Media Player Classic Home Cinema (MPC HC), XBMC 12.0 Frodo.

ZOTAC ZBOX nano VD01 play youtube 1080p vs. Ulrabook and Mininotebook
http://www.youtube.com/watch?v=st05cMVegSU
• H.264 AVC 1080p DXVA HW acceleration: max. 24W (90% of time is spent on 23W)

VIA VE-900 Mini-ITX Desktop Board power consumption idle (VIA Nano X2 L4050 1.4GHz)
idle: 16W !!! (• Akasa CryptoMini ITX 60W)
source: http://www.youtube.com/watch?v=_rr9iV056s4
• High-performance 64-bit VIA Nano X2 L4050 1.4GHz dual core processor 800MHz V4 bus
• VIA Chrome9™ HD IGP Hardware acceleration for VC1, H.264 and WMV9 HD video
• On board HDMI and VGA ports
• Support for up to 8GB of DDR3 1066 DIMM (2x 4GB Kingston HyperX DDR3 1600MHz CL9@1066MHz CL7)
• Support for 2 SATA (Kingston SSDnow V200 64GB)
• PCI slot expansions slot
• Akasa CryptoMini ITX 60W


Please. How much exidence (argument) do you need?

So what you're saying is that even with a newer southbridge, it still has worse idle power than Bobcat? And even using hardware acceleration, it still has worse video playback power consumption than Bobcat? And I notice you've failed to address that the Prime95 power consumption was almost double that of Bobcat.