|
I never really saw the need for more than two cores, as there until recently we're to few useful programs that could use multiple cores.
Well, there are games, which would ruin what little I have left of a social life. And databases, which I get enough of at work.
I do have some hopes for graphic programs, but Amdahl's law makes me suspect that I will never buy a personal computer with more than four cores.
|
|
|
|
|
| Moore's law and process shrinks would argue that in a decade or so you'll probably be unable to buy one with that few. Chip makers are confident that they can push silcon to the 11nm process node. That's ~16x transistor density of the current mature 45nm process (32nm capacity is too volume limited at present to be counted as mature). At 45nm we have 1-4 cores depending on your performance slot. Handwavingly that would imply 16-64 cores on 11nm chips. Even though the CPU's currrent habit of eating chipset functionality will probably degrade that number on the lowend I suspect we'll end up with x86 chips with 8 cores as the minimum. If the 7 or 8nm node turns out to be possible as well the number will go up. If none of Germanium/Gallium Arsenide/Gallium Nitride/Graphene reach a level suitable for mass market CPUs before Silicon maxes out things will get interesting... 3x12=36
2x12=24
1x12=12
0x12=18
|
|
|
|
|
You're quite right.
And the fact that most programs (in the near future atleast) will continue to do most of their work in just one thread, doesn't change the fact that we will have a raising number of programs running in the background.
I can only blame that I have a single threaded mind and that it's late. 
|
|
|
|
|
Jörgen Andersson wrote: I can only blame that I have a single threaded mind and that it's late. Smile
Don't feel bad. It happens to the best of us. 3x12=36
2x12=24
1x12=12
0x12=18
|
|
|
|
|
Jörgen Andersson wrote: I will never buy a personal computer with more than four cores.
That reminds me of a Bill Gates who once said no computer would ever need more than 640 kilobytes of memory.
Dan is correct, the minimum will rise.
BTW: a PC meant to be a server may lack a video card (can be added), and some audio circuitry; you might not be able to get the elementary beep client versions of Windows sometimes want to throw at you.
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
I only read code that is properly formatted, adding PRE tags is the easiest way to obtain that. All Toronto weekends should be extremely wet until we get it automated in regular forums, not just QA.
|
|
|
|
|
Note to self:
Never use the word never!
|
|
|
|
|
... unless there is another negation in your sentence.
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
I only read code that is properly formatted, adding PRE tags is the easiest way to obtain that. All Toronto weekends should be extremely wet until we get it automated in regular forums, not just QA.
|
|
|
|
|
I personally have wanted to get a desktop with Nvidia Tesla cards, but I do a lot of computation.
|
|
|
|
|
I've used dual socket dell workstations and single socket dell test machines since 2005. In my experience, the operating system of the dual sockets hung a lot less than single sockets with the same cpus. My guess is that you could recover from infinite loops and other run-time errors from poorly designed anti-virus, anti-spyware, OS, etc. in dual socket, but not single socket.
|
|
|
|
|
Hello
I am a newbie with DSF and USB
I looked at the MSDN sample for simulating a USB Keyboard (generic HID) using DSF.
I was able to run the sample and have the keyboard enumerate and simulated device driver loaded but was not clear on how to construct and send input reports to the simulated device using the provided script file.
Further I tried to modify the script to simulate a mouse and again was unsure of how to communicate with the mouse.
My end goal is to be able to command the simulated mouse cursor to scroll on the screen.
Any help is highly appreciatedSreedhar
|
|
|
|
|
Try my sample code from my kinect project at: http://kinectmultipoint.codeplex.com
It simulates mouse input but the dsf script needs to be converted to vb.net
jeffery
modified 11-Jan-12 13:27pm.
|
|
|
|
|
Mum got me the zen xfi2 for christmas which i'm pretty happy with till today. And today I got to know that creative has released the adk (application development kit) for it. Take a look at www.creative.com/zenxfi2/adk However, i'm not good at programming so was wondering if anyone here has experiences with Lua or ADK? If yes, pls see if you can help create more apps (like games) for it. That would be really useful
|
|
|
|
|
I am a embedded system program currently working on microchip PIC18f4550
controller which is useful for USB communication.
I had search on Microchip web site but not get it.
Any help/link appreciated
Khaniya Sunil
|
|
|
|
|
when I enter microchip productid into the Google search box, I get a lot, starting with this[^].
Does that help?
Luc Pattyn [Forum Guidelines] [Why QA sucks] [My Articles]
I only read code that is properly formatted, adding PRE tags is the easiest way to obtain that. [The QA section does it automatically now, I hope we soon get it on regular forums as well]
|
|
|
|
|
Thank you very much for your response
I have already found it. Microchip controller PIC18F4550
VID (Vendor ID) = 0x04D8
PID (Product ID) = 0x0011
Thanks agin
|
|
|
|
|
I have bought FT232BQ USB UART IC , it is mentioned in the data sheet that it supports Isynchronous data transfer mode with an option bit in the EEPROM , but i cant understand how to do so..
what is the max. data transfer rate i could achieve using this IC and how can i acheive it ,, thnx in advance
|
|
|
|
|
ammeer_a wrote: max. data transfer rate i could achieve
isnt that (from the datasheet) "A new EEPROM based option allows the FT232BQ to return a USB 2.0 device descriptor as opposed to USB 1.1. Note : The device would be a USB 2.0 Full Speed device (12Mb/s) as opposed to a USB 2.0 High Speed device (480Mb/s)."
'g'
|
|
|
|
|
yes , but this maximum speed can only be achieved in Isychronous data transfer , and i dunt know how to get this rate ..
|
|
|
|
|
i have made a simple hardware interface consisting of an ADC and am trying to connect it to my labtop through usb,,i searched the net and i found that to access usb i need to have a driver for the device and this is done in kernel mode , unfortunately ,i dunt know how to write a driver although am familiar to c++ and vb codes
if any body can help me to write this driver ,, I need it to allow me to send and recieve bits from the ADC (in serial of course)
|
|
|
|
|
I did a similar thing about ten years ago though I bought an empty PCI card for my circuitry rather than using USB. When I realised I needed to write a driver I looked around and ended up buying a development kit from VToolsD which allowed me to program a driver, based around a sample, using C++.
|
|
|
|
|
Whats ADC?
Wouldnt you be better connecting it to the serial port? You should be able to use the onboard serial driver. To create a USB device you need a lot of hardware first, adn then writing a driver will probably take you about 6 months plus.
Dont underestimate the kernel or attempt to create a driver unless you are seriously committed. There are a lot of pitfalls and it takes a long tine to learn it properly. Morality is indistinguishable from social proscription
|
|
|
|
|
I've got a RAID project to run into.
As you know, MS disk manager provides soft RAID managment such as creating RAID disk array and reconstructing disk image on RAID0 RAID1 RAID5.
Is there any API or DEVICEIOCONTROL code in C++ so I can manage RAID in my APP instead?
modified on Sunday, January 31, 2010 10:36 PM
|
|
|
|
|
dont have idea about APP but if you wanna configure raid in your pc i might be able to help you
|
|
|
|
|
How often should one defrag one's hard disk (or partitions)? Is defragmentation really useful?
|
|
|
|
|
The answer is simply, it depends.
I'm running automatic defragmentation every week for our fileservers and databases.
A drive that's used only for reading don't need it.
A workstation ends up somewhere inbetween.
There are several points with defragmentation:
Reading a file from the disk is faster if it's contiguous. (unless you have a solid state drive)
Less file segments takes less space in the MFT which might make the drive a bit faster.
In the case of a harddisk crash the chance that a recovery tool will be able to save your data is greatly enhanced if the files are contiguous.
Overusing defragmentation is of course pointless as it shortens the life of the disk and everything gets slow during the actual defragmentation.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
