Trouble with dual graphics is that the R7 250 is around the same price on the street as the R7 250X (actually R7 250X is actually often times cheaper), but the R7 250X is faster by itself than Kaveri iGPU + R7 250 together by a considerable margin.
It seems that they gave up this sales bullet. Pity, it would give them a real edge on gaming.
Trouble with dual graphics is that the R7 250 is around the same price on the street as the R7 250X (actually R7 250X is actually often times cheaper), but the R7 250X is faster by itself than Kaveri iGPU + R7 250 together by a considerable margin.
I plan on just re-running the settings reported in this post:
http://forums.anandtech.com/showpost.php?p=36730608&postcount=133
(Except instead of dual channel DDR3 1333, I will run single channel DDR3 1333....and maybe single channel DDR3 1600 as well. Just realize Pentium doesn't support DDR3 1600 on Non-Z boards, but I have a Z97 so I can do it for academic reasons).
Besides some legacy compatibility issues, is there any difference between memory allocated in BIOS and the dynamic memory allocated by the graphics driver?With regard to increasing RAM allocated to iGPU, I did find out my board cannot allocate more than 512MB to iGPU
I did find Beema Laptops (during my newegg search), but Yes Ancalagon44 the selection was limited to models with screen size 15.6" or greater. I didn't see any Mullins Laptops.
Because id rather have a thuban than anything AMD has released recently. Their CPU performance is dismal at best.
If my FX's have been dismal I hope to hell every other computer I ever have is equally so lol
Come to the dark side, we have cookies!
Sometimes I wonder what kind of cpu workload the average AT-ers would have. 'cos at this point I dont really care about cpu speed anymore. My craptacular (compared to Intel's latest) Kaveri does fine even when compiling a program in Codeblocks, having a virtualbox machine running, a couple of AWS ssh and with tons of tabs open in Tor-firefox, normal firefox and Chrome all at the same time. Its not like I stare at the screen waiting for 7-zip or g++ to finish. But then again, I dont play the latest AAA titles anymore.
Here are my Dirt 3 results (the results posted in the link above were 2 x 2GB dual channel DDR3 1333):
3 GHz Pentium with 4GB RAM (single channel DDR3 1333), 512 MB of RAM dedicated to iGPU
Ultra Low pre-set (with Multisampling OFF):
1920x1080 = 27.74 FPS, 24.59 MIN
1366x768 = 45.15, 37.56 MIN
1024x768= 55.29 FPS, 44.25 MIN
800x600 = 69.85 FPS, 53.46 MIN
Low pre-set (with Multisampling OFF):
1920x1080 = 21.11 FPS, 17.43 MIN
1366x768 = 36.01 FPS, 30.21 MIN
1024x768 = 43.22 FPS, 34.47 MIN
800x600 = 52.13 FPS, 39.67 MIN
Medium pre-set (with Multisampling OFF):
1280x720 =
AVG: 29.94 FPS
Min: 25.51 FPS
(Comparing my single channel results to the dual channel results in link, I am quite surprised to see the FPS drop so much.)
One problem with Kaveri was that is was supposed to have GDDR5. It also has a jumbo size DDR3 PHY that is twice the size of the Trinity APU DDR3 PHY (I am thinking the GDDR5 is the one reason for that large PHY).
Anyway, in retrospect, AMD could have designed a quad core steamroller based APU with small iGPU and reduced sized DDR3 PHY (with the GDDR5 removed) for about the half the size of the current Kaveri.
Why do you keep bolding TDP?
To emphasize the importance of the low TDP.
The dominant thread in a game, and maybe Firefox. That's really it. At the right price, Kaveris are awesome, because the generational improvements in CPUs, and Turbo (which I was against when it came out), make more of a difference than anything else, for day to day bursty CPU loads, than the higher IPC does (but, the higher IPC is one of several things that allows Intel to offer lower power consumption, which sells to big customers, so let's hope for something good in 2016 from AMD!).Sometimes I wonder what kind of cpu workload the average AT-ers would have. 'cos at this point I dont really care about cpu speed anymore. My craptacular (compared to Intel's latest) Kaveri does fine even when compiling a program in Codeblocks, having a virtualbox machine running, a couple of AWS ssh and with tons of tabs open in Tor-firefox, normal firefox and Chrome all at the same time. Its not like I stare at the screen waiting for 7-zip or g++ to finish. But then again, I dont play the latest AAA titles anymore.
Small Intel boards had that, too, before Haswell (or they had heatsinks on 4-phase VRMs, and needed them badly). Without the voltage control on the CPU, the mobo's VRMs have to adjust to every large change in current on the chip. TDP is a measure over a long period of time, while the voltage regulation has to be designed for peak instantaneous (well, very short-term) needs, which have not been decreasing on the desktop. I suspect the i3 being sufficient (a better GPU means nothing if you don't need it), sometimes better, and definitely more popular, has more to do with it than technical difficulties, though some of the cost may still reflect additional BOM costs.Im not sure about it, something is really wrong with FM2/FM2+ overall, for having such low tdp its kinda strange that there are soooo little FM2/+ Itx boards, and the ones that are avalible are expensive and they come with a crazy number of vrms and most of them have a heat spreader too.
I smell something wrong with it, one whould think that ITX will be the main target of FM2 platform, and it just does not exist on it.
So yeah the tdp is low, but something is definately wrong with power or heat on FM2.