Print 79 comment(s) - last by ceefka.. on Jun 19 at 11:17 AM

Mockup USB 3.0 cable  (Source: Tech On)
Intel says open host controller specifications have cost gazillions of dollars to develop

According to Intel’s Nick Knupffer, there are a lot of myths going around concerning USB 3.0 and Intel’s involvement in the development of the specification. Knupffer wrote a blog post on Intel’s website in an attempt to dispel these myths.

Knupffer points out that Intel is not developing the USB 3.0 specification. What Intel is developing is the host controller spec which Knupffer describes as a “Dummies Guide” to building a USB 3.0 compatible piece of silicon.

Knupffer says in the blog post that Intel has invested “gazillions of dollars and bazillions of engineering man hours” in developing the open host controller and despite its significant investment still plans to give the specification to competing manufacturers for free. Knupffer also says that Intel loves it when CPU performance is used to the max and the huge increase in bandwidth of USB 3.0 will mean larger file transfers and more processor usage. This in turn is expected to lead to an increased demand for faster processors.

AMD and NVIDIA leveled allegations at Intel recently that claim Intel was withholding the open host controller specifications in an attempt to give itself a market advantage. Intel and AMD claim that by withholding the specification the lead Intel will have in bringing USB 3.0 compliant products to market will be in the six to nine month range.

Intel denied the allegations of withholding the open host controller specifications at the time AMD and NVIDIA made their charges public and announced they would be designing their own open host controller. In Knupffer’s blog post, he again says that Intel isn’t holding the open host controller specifications back from competitors.

According to Knupffer, the significant investment in the open host controller specifications is specifically to get USB 3.0 into the market faster, so why would it withhold the specification. Intel still maintains that the specifications aren’t ready and that it plans to give the specifications to other manufacturers in the second half of 2008.

The final myth that Knupffer addresses in his post is that USB 3.0 technology borrows heavily from technology used in PCI Express. Intel points out that it was involved with both the PCI-SIG and the USB-IF at the design stage for both PCI Express and for USB 3.0. The insinuation form Intel is that the technology that is similar in both devices was developed on its dime.

Comments     Threshold

This article is over a month old, voting and posting comments is disabled

RE: USB processor
By mattclary on 6/12/2008 2:34:18 PM , Rating: 2
I was wondering why no one else seemed to pay attention to that quote. It is my understanding that Firewire is a lot less processor intensive. All of a sudden, the processor demands of USB seem to make sense, don't they?!

Really kind of ticks me off that they want to design stuff to make demands on the CPU when it might be avoided.

RE: USB processor
By Reclaimer77 on 6/12/2008 3:23:23 PM , Rating: 1
Really kind of ticks me off that they want to design stuff to make demands on the CPU when it might be avoided.

Why not ? How many times is your CPU at 100% loads or even 50% ? My " old " E6400 is at 15% load right now with Firefox open, Winamp running, a virus scanner and two spyware apps going AND 43 processes running. And the second core is barely doing anything at all.

Why not utilize the CPU more ? There is so much potential we're not even exploring. This is a good thing Intel is doing !

RE: USB processor
By mattclary on 6/12/2008 3:31:53 PM , Rating: 2
If CPU usage doesn't slow down the transfer or cause other processes to slow down fine, but in my experience, even somewhat fast machines can be bogged down during a USB transfer of large files or a large number of files.

RE: USB processor
By mattclary on 6/12/2008 3:41:38 PM , Rating: 2
Do a google search on: usb cpu usage

I think a big part of the problem is being CPU bound, poor driver implementations can wreak havoc causing pegged CPU usage.

RE: USB processor
By SiliconAddict on 6/13/2008 1:30:36 AM , Rating: 3
Yah and lets say I'm ripping\encoding a DVD directly to an external drive those clocks are being eaten up as it accesses USB 3. Pointless and stupid. FW800 may be expensive but I have no problems treating attached drives as internal drives because of the speed.

RE: USB processor
By Klober on 6/13/2008 2:59:35 PM , Rating: 2
Why not utilize the CPU more ? There is so much potential we're not even exploring. This is a good thing Intel is doing !

Let's think about this for a moment. Intel says that everything they do now takes power consumption into account, yet this shows that as a flat-out lie at worst, and misleading at best. The less the CPU is in use the more chance it has to go into a low-power state to save energy, but by running the USB bus off the CPU you are guaranteed that as long as you're doing something as simple as moving the mouse you're going to be using CPU cycles. Also, a specialized processor has the capability of being more efficient than a general purpose processor, yet they keep the USB relying on the CPU instead of giving the USB interface its own processor. This means that you are going to use more energy to run the USB, which in turn produces more heat, which in turns requires more energy to cool off the room it is in, all of which is a continuing higher cost to me. I would prefer, for the reasons stated here along with several others (i.e. use my CPU to run AI/physics/etc for Crysis, instead of ALSO helping run the USB to process my keyboard and mouse commands and whatever else is attached), to spend a bit more on the initial one-time cost and keep the recurring costs to a minimum. And I'm pretty sure building a bit extra onto the chipset to process USB signals, or creating a small chip specifically for that purpose, is going to be more cost-effective for the end-user in the long run, not to mention it'll help with the coming (already starting) energy crisis.

Just my 2 cents.

RE: USB processor
By theapparition on 6/12/2008 4:36:58 PM , Rating: 5
It is my understanding that Firewire is a lot less processor intensive.

Your understanding is correct, but interpretation is all wrong.
USB is a LOW COST interconnect that uses the processor to manage communication. Implementation of USB into integrated chipsets is soooo low cost, it's almost not even worth mentioning.

Firewire (IEEE1394) has it's own logic to control comms. As such, it's implementation costs signifigantly more. It did have the added benefit of not loading the computer as much.

As such, both formats have co-existed and will both continue to evolve. You see this clearly diferentiated between target markets. Where devices exist that need high speed serial communication between devices where a computer isn't present, you'll only see Firewire. TV's, Camcorders, STB's, etc. All Firewire because a CPU isn't required. Where computers are involved, you'll see mainly USB.

See, there is reason behind the madness and not all is for "nefarious motives".

RE: USB processor
By mattclary on 6/13/2008 9:39:09 AM , Rating: 2
Thanks for the response, that makes a lot more sense to me now.

"I want people to see my movies in the best formats possible. For [Paramount] to deny people who have Blu-ray sucks!" -- Movie Director Michael Bay
Related Articles

Copyright 2016 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki