And seems like you haven't read about fusion APU. It's chip with CPU+ DX11,OpenGL 4.1 GPU in a single package.
I didn't read about APU. anyway according to what I know DX11 and OpenGL are just libraries.And they can use some
hardware support indirectly[they are not standalone with the software layers]. Todays GPU cards are not build like that. But in older days they did made as the way
you have spoken. Did you remind 3dfx cards, [that time I was at grade 7 or 8 as I could remember, and my father
have worked with a shatted wire-frame graphics package , had to wrote tons of macros when you have to
do a simple perspective projection of a simple model].Glide was not work as a library , but as a core
hardware platform when we talk about 3dfx cards. But the 3dfx time is dead due to lots of reasons now.
unlike OpenGL/DirectX which have a hardware abstraction layer 3dfx Glide only have some routines/macros
that help to avoid programer doing dangours things like chip initialization.There was no HAL in Glide.
I just read few chapters randomly from GPU GEMS[there are 6 volumes , I'll upload if you can't find them] , but me also not processing a computer graphics professional
industry level skills. please correct me if I'm wrong.But I have worked with those 3dfx cards !
In the Gforce 4 times , hardware T&L is built into the GPU,so a library like OpenGL can directly transfer matrix
math [transformations] and Lightning into the hardware level. But it still we can't say DX11 or OpenGL4.1 is a
hardware feature.It's just have supportive hardware, and there always a HAL, hardware abstraction layer unlike Glide times[that Glide time is the old but a golden /at lest silver time in GPU computing] I'm sure you wrong in that facts.
And pixel shader and vector shader hardare support have been added to when it comes to DX8, and I still didn't
dig enough to DX8 and above to talk about them. But one thing is sure DX11 never can be standalone without a
HAL ,OpenGL also, because DX8 have a HAL , which means it's necessity to DX11 card to have a HAL.
Seems like you haven't heard about Sandy Bridge CPUs and Intel HD graphics.
I didn't go through the Sandy Bridge CPU. Even they name the new microarchitecture as Sandy Bridge/Ivy Bridge,
the things like GPU and craming GPU into the CPU die doesn't mean it's belongs to the core. Core/microarchitecture belongs to these concepts.
*interprocessor communication /HT/threading/locking
*caching
*pipelining
*Branch Prediction(intel have spend lots of research on this)/super scalar.
..etc.
QPI is just the technology which replaces the Font Side Bus technology, so it's not belongs to the
microarchitecture.You can say it's belongs to the microarchitecture because wiki page is mentioning
QPI under the Intel Core and Nethelm microarchitectures , but at that time i read that page,info
are most outdated. Go for more clarifications....and there is no hard and fast margine which isolate the
microarchitecture.
The changes to the on DIE GPU was came with sandy bridge, but still it's not belongs to sandy bridge
microarchitecture which is developed in Israel[jews and Israel , my favourate community and country].
Please clarify and understand clearly about the terms die, core, package very clearly.
here: [just googled for ya]
http://superuser.com/questions/324284/w ... nd-package
That means, memory controller, PCI-e lanes are inside the CPU. This is to reduce the latency. As i can remember, AMD was the first to do this. Quick Path Interconnect for Intel and Hyper transport for AMD are processor to NB connecting technology.
ya, but still it doesn't belongs to microarchitecture, what belongs to the microarchitecture is what's inside a core.
and communications between cores, locking and synchronization etc.
About fusion I don't know to talk about their microarchitecture in AMD, I never read a AMD system programmer's
manual since 2005.[the funeral of my AMD XP2600+ processor].But facts should be same ,they would not be
belongs to microarchitecture, but it can designed inside one die [one silicon pice] to reduce latencies.
They are very low power but still powerful chips that can run 1080p videos without any problem at all.
is that Intel Quick Sync? Never read it anyway. But intel is promising you can view a blue ray HD video with
just their integrated graphics chips , isn't that fake?
![Laughing :laughing:](./images/smilies/laughing.gif)
, true my i3 with Nvidia 320M GPU still
flickering when it plays a movie from a external HD Blue Ray.I don't know may be USB speed is not enough.
and even not think about that kind of super technology yet.
I can't comment on socket changes. I do not have any idea on that. If there are lots of power fluctuations in your area, buy a good power stabilizer and a good power supply unit with at least passive PFC.A PSU unit with that specs is about $100 or more in local market.
why you think they change it for a marketing reason? Even in the same microarchitecture they need to
change the pins of the package when they moving to higher speeds. Because inductance and capacitance
and therm-dynamic behavior/characteristics of processor made it necessary. Find out what's the most
valuable traces of a typical motherboard? Since QPI like technology is so dynamic , if you keep your motherboard
clean it will open more paths[virtual] to it's peers dynamically.
and about a passive filter, I want to apologize it's completely my mistake on own wiring on my room. thanks to
my own carelessness .Even after I have a filtered power supply it won't help either.you come and see my room by
yourself.
If everyone is going to buy intel, there will be no competition in the market. That is bad for the consumer.
![Wink ;)](./images/smilies/icon_e_wink.gif)
.
I'm not a sales expert[who have sold 1000 fireballs to 10 customers] but when we talk about company price,
it have nothing to do with local sri lanka markert domination. because they declare prices to the international
market. If there are more than one dealer for Intel products then there will be a good competition. No need
a separate brand for that isn't it?
For me, it's a pain to use my laptop.
let me know why?
I do not have a particular hate for Intel. But buying a good motherboard for AMD platform and a good CPU dose not cost a lot like for Intel. I am not talking about cheap Foxconn motherboards. I am talking about real motherboard from manufacturers like Gigabyte, Asus, MSI etc which offers great features,stability, support and value for money.
Nobody hates intel or AMD. If they offer you a job opportunity do you say no? They are technologically
superior companies.The bad thing is how they plan their sales strategy. About Foxconn, it's cheap because
it's assembly lines are located in China slave camps
http://www.telegraph.co.uk/finance/chin ... ctory.html I don't have any intention to boycott foxconn but facts are
facts , so I decided to leave that above link.
and about Asus , i really dreamed to buy a Asus laptop, but it's too costy then i went for a ACER 4741.
AMD might produce lots of heat. I do not have any idea on that.
that's due to their technological weakness. On the other hand intel uses something called high-k meterial.
http://en.wikipedia.org/wiki/High-k_dielectric
[don't trust the wiki links more, I just put the google's first search for ya anyway].
so they can reduce their leak current. that way they can reduce the heat. and 32nm technology, the more
it smaller the more leak current is. There is something called power wall, [in sem III you'll learn these for sure],
and manufactures not exceed that power wall but AMD did, and it depends on the external cooling mechanism.
and with hi-k hi dielectric materials can built smaller transistors without exceeding that power wall.
take for a example core2 duo and i3 , sales ppl will told you the only difference is HT was implemented in i3
but not in core2duo, but another thing they have to say, it's power consumption was reduced by that
above mentioned new technology and smaller transistors compared to core2duo.I can really feel that , my acer
i3 lappy heats very very low.