You mistake criticism of the initial mobile efforts, which was still justified IMO, as company hate which is too strong and incorrect an association. I made clear at the time that I still really liked the Core lineage right from its Pentium-M roots to the present day. I also did not care too much from the Pentium 4.
It really was not a given until it was proven and proven well with Medfield because 45nm Moorestown had a nasty power spike at high clockspeeds which I pointed out from its datasheet. Although x86 is just fine for mobile chips current implementations by everyone suggest that RISC is still better suited for 1mW 100 MHz micro-controllers. Quark is a pure 486 shrink and is higher than that but it really does not matter much because the other board items like DRAM will drown out its contributions.
So the x86 power inefficiency compared to RISC is probably true to a certain extent but only at a very low level no-one really cares about for application processors and at 1W+ levels the x86 CISC nature helps it pull ahead in performance per issue width which equates to performance/power in the end.
Don't be too hard on Merrifield, consider it an essential part of the Atom family which is efficiently Medfield-sized and will prove very useful cost-wise when it is relegated to the middle and lower tiers after the 14nm Airmont and Goldmont generations arrives. There will be plenty of time for barnstorming phone quad-cores at 14nm.
Before Medfield and Clover Trail Intel's Atom efforts were so incredibly poor it was hard to believe otherwise or that the architecture could be rescued in any way but it was. In a way it's good that the netbook versions of Atom have so convincingly died out now so as not to leave a bad lasting initial impression of Atom. Medfield/Clover Trail(+) will stick around now for some time in cheap slots to erase any bad memories of the first generation Atom.
I kind of predicted the up/down path the Tegra family would take up to now with all the intense competition it would face. However going forward I can see enough differentiation on the gpu side and the device side (e.g Shield/Note families) to see that this could perhaps be a viable profitable business sometime in the future. Nvidia is never going to take down the giants of this industry but it does not need to, it just needs to make a bit more money for its shareholders. Huang also has made good business connections with HP and Chinese OEMs which will also help in selling his hardware.
The gpu side though is problematic with the hard competition from AMD and Intel but the very high-end keeps rescuing it from collapse but it does look like a bit of a tough treadmill.
If revenues/earnings are flattish as they are, by reducing the share count you can still keep on increasing the eps and extra dividend always helps to placate the market. Huang has played his not so good cards well for the shareholders. The stock also finished over the 50 MA of 15.41 meaning the assault on the long term moving average resistance in the 16s can resume.
I'm with the mas guy on this ;-)
mas Friday, 11/08/13 02:39:16 PM
Re: aeassa post# 124738
Post # of 124744
50+% is multi-threaded performance and by the same reasoning Cyclone is only a Snapdragon 400/Tegra 4i competitor. You know this already so why you continue to be unduly negative knowing this is beyond me. If two high performing cores with a PowerVR6 gpu is good enough for Apple why is it not good enough for Intel ?
No problem. These quad-core A7s are the devices that Clover Trail+ should be taking out and the battery life extension is only due to using large notebook sized batteries. Still Intel should be going after and getting these sockets.
He already knew that which is why he purposely posted it in a Mediatek/Lenovo thread. Are you naturally this dumb or do you have to practice ? It's called scouting the opposition.
There is medium term moving average support (~380-450 DMA) at 83c. There is long term moving average resistance (~ 200-230 WMA) that all peaks at 88c. If 83c breaks it will likely keep falling. If 88c is breached it will likely keep rising otherwise it will remain in the tight 83-88c band until further news affects the stock.
'ARMH looks broke to me ever since their last earnings report. The volume and buying indicators have tapered off and on friday it spent about 15 mins below its 50 MA but recovered just before the bell. It's almost like the smart big money has left the building and the bagholders have not got the message yet. It's definitely drifting down and beginning to test all its technical support.'
Well it sure broke well and good today. It absolutely crashed through the 50 MA of 46.93 and the next moving average support is at 43.22-43.28 where the 200 and 100 MA are hanging out. Stock Consultant says it is currently resting on a 45.1 ± 0.72, type Double, strength 4 Support line but below that the next fixed support is far away at 41.67 ± 0.67, type Triple, strength 7 which is below the 100-200 MA band. The likelihood of ARMH now dropping into the 43s is very high and the last time it fell through its 50 MA in June it did not stop dropping until 35. Humpty is very seriously sick now.
Ashraf is right, Geekbech is next to useless and all the x86 FP bugs shown up in it recently in Version 2 that the writer did not even know he had before it was pointed out to him show how clueless he is.
SpecCPU is the gold standard in benchmarking and anyone can neutralize any compiler optimized differences by choosing the same compiler to test different architectures with in the same review, computerbase do this. ARM, Intel and Nvidia agree with me and have given Spec2000 estimates for their mobile chips with Intel also adding Spec2006. The fact is that SpecCPU tests all the parts of a processor core/memory architecture not just the core which is what all your favorite tiny micro-benchmarks do. Spec2000 also runs short enough to test peak mobile performance while Spec2006 runs long enough to test sustained mobile performance so that would be another benefit. These review sites have got to stop being lazy and run SpecCPU themselves rather than these toy benchmarks to get a truer idea of mobile processor performance.
I don't. The last time the GSE backlog was specifically mentioned in the SEC filings it was $9m last year so if none had been booked until last quarter GSE revenue could have been as high #$%$7m last quarter.
INTC is being strongly supported by all its short and medium term moving averages between 22 and 24. These are all rising and causing upward pressure on the very long term moving average resistance between 24 and 26. Over time and given no bad news they will take out this resistance and the stock will then move much more easily in an upward direction.