Intel Haswell to arrive in low volume until USB 3.0 issue is fixed

Hilbert Hagedoorn

Intel will slowdown the ramp of its new Haswell platform until it has fixed a USB 3.0 issue in the Haswell platform that causes some relatively minor issues when a system wakes from S3 sleep mode with...

Intel Haswell to arrive in low volume until USB 3.0 issue is fixed

nicugoalkeper · 03-15-2013, 09:19 | posts: 737 | Location: RO TM

Or another way to say we are selling faulty chips !!

hallryu · 03-15-2013, 09:22 | posts: 10,888 | Location: England

I will part with any cash until the problems are fully resolved! I need a new case too!

NAMEk · 03-15-2013, 10:45 | posts: 189 | Location: Lithuania

Disgusting.

RoadKillNL · 03-15-2013, 10:55 | posts: 2,679 | Location: Netherlands

Bah, guess i'm gonna have to wait even longer.

warlord · 03-15-2013, 12:11 | posts: 374

help me all these from the beginning problems, can't be fixed with just a bios update.? if not, why.?

dsbig · 03-15-2013, 12:22 | posts: 1,739 | Location: in a house

I dont even use my usb 3, it disabled in bios.

speeds up my boot process when it disabled, so even if I had a haswell chip that had the issue it wouldnt bother me.

I dont even use sleep mode.

{LOD}PITBULL · 03-15-2013, 13:11 | posts: 768 | Location: Upstate NY

Hi WE are Intel WE are going to over charge you and you are going to PAY...lol. But wait we also Put out Faulty chips which you will Also pay for.. Suckers... lol.

Denial · 03-15-2013, 13:17 | posts: 6,374 | Location: Above Earth in a Big Rocket Ship

I hate to break it to you guys but every single processor ever made has some form of errata. I mean yeah this is more serious obviously, but if you think they haven't' sold you "faulty" chips in the past you're sorely mistaken.

dsbig · 03-15-2013, 13:18 | posts: 1,739 | Location: in a house

companies do it...

I had compaq computer that wouldnt actually save power in sleep mode because the psu where cheap and didnt have sleep mode. the motherboard would go into sleep mode, but the psu stayed on.

Nono06 · **Memory**: 6GB (8-9-8-24-1T)@1870

Just a stupid question, how many people here use PC sleep mode/resume with USB 3 drive connected and used by an application while going to sleep?

Moreover, just restarting the applications or unplug replug the drive could do the trick. For me it is clearly a minor issue since I only have Sata external drives

-Tj- · 03-15-2013, 14:13 | posts: 2,667 | Location: Downtown ✈

Im going to get it anyway, i dont use usb3 and hardly put cpu to sleep so its not a big deal for me.

Dragondale13 · 03-15-2013, 14:27 | posts: 90 | Location: Trinidad & Tobago, W.I.

Minor issue, I'm holding out cuz I wanna see NUMBERS, NUMBERS!!!! I LIKE NUMBERS!!!!

schmidtbag · 03-15-2013, 14:31 | posts: 112 | Location: MA, USA

Quote:

Originally Posted by Denial

I hate to break it to you guys but every single processor ever made has some form of errata. I mean yeah this is more serious obviously, but if you think they haven't' sold you "faulty" chips in the past you're sorely mistaken.

Based on the way you're wording that, fault is a relative, maybe even opinionated term. For example, the Phenom II x6 can sometimes outperform an 8-core FX processor, where you could consider the FX to be at fault when really it just isn't optimized for that particular task; the FX isn't worse than the Phenom. In a situation like this, a more suitable word would be deficient or regression.

So, perhaps what you meant to say is every chip ever sold has deficiencies/regressions, because it is simply un-realistic to have 1 chip be good at everything.

If you really do mean that every chip has had a fault, what were the fault(s) of Sandy Bridge or Core 2 Duo or Athlon II? I'm not saying those didn't have faults, but I simply haven't heard of any. And by faults I mean in ways the chip was intended to be used, so faults in overclocking don't count.

Ghosty · 03-15-2013, 14:55 | posts: 1,454 | Location: UK

Will the Haswell CPU's work with existing X79 chipsets?

Denial · *Last edited by Denial; 03-15-2013 at 15:00.*

Quote:

Originally Posted by schmidtbag

Based on the way you're wording that, fault is a relative, maybe even opinionated term. For example, the Phenom II x6 can sometimes outperform an 8-core FX processor, where you could consider the FX to be at fault when really it just isn't optimized for that particular task; the FX isn't worse than the Phenom. In a situation like this, a more suitable word would be deficient or regression.

So, perhaps what you meant to say is every chip ever sold has deficiencies/regressions, because it is simply un-realistic to have 1 chip be good at everything.

If you really do mean that every chip has had a fault, what were the fault(s) of Sandy Bridge or Core 2 Duo or Athlon II? I'm not saying those didn't have faults, but I simply haven't heard of any. And by faults I mean in ways the chip was intended to be used, so faults in overclocking don't count.

I'm not talking about performance characteristics, I'm talking about processor errata. Intel documents it on the release of every stepping. Certain programming functions will literally be broken on certain processors. For instance Intel had a TLB bug in it's MMU on Core 2 which would cause hardlock ups of the processor when performing certain operations.

Here is an example of the documentation that Intel releases for programming around errata:

http://download.intel.com/design/mob...T/31407918.pdf

If you scroll down to page 40 you can see the list of all the problems and workarounds for them.

http://www.anandtech.com/show/2477/2

Anandtech did its own write up on the AMD version of the TLB errata.

Most of these problems aren't seen by end users because developers are aware and make changes in order to fix, or they are fixable via bios/microcode updates. This USB3 problem I guess is more significant and cannot be fixed via these methods but by a new stepping. But to pretend like processors never have problems and Intel is just now selling a faulty chip is nonsense.

freeZ · 03-15-2013, 15:03 | posts: 658

Quote:

Originally Posted by Nono06

Just a stupid question, how many people here use PC sleep mode/resume with USB 3 drive connected and used by an application while going to sleep?

Moreover, just restarting the applications or unplug replug the drive could do the trick. For me it is clearly a minor issue since I only have Sata external drives

I do. Seems pretty inconvenient for the bleeding edge. Think I'd rather wait till they ironed out this problem before considering it.

RoadKillNL · 03-15-2013, 15:47 | posts: 2,679 | Location: Netherlands

Quote:

Originally Posted by -Tj-

Im going to get it anyway, i dont use usb3 and hardly put cpu to sleep so its not a big deal for me.

You're right.

I'm not using Sleep Mode either, nor USB 3.0 on my current motherboard. (But that's because i can't route them to my front panel) I do have a USB 3.0 External HDD and Thumb Drive tho, but i'm a patient person

I think ill just go for Haswell as soon as they are out. By that time my current CPU hits 3,5 years of age, it's time to retire.

RoadKillNL · 03-15-2013, 15:49 | posts: 2,679 | Location: Netherlands

Quote:

Originally Posted by Ghosty

Will the Haswell CPU's work with existing X79 chipsets?

Haswell has 1150 pins, x79 has 2011 pins.

The Chubu · 03-15-2013, 16:11 | posts: 2,541 | Location: Look out!

Intel: "Haswell is going to be relased... sometime. It's not like we have to check our release dates against AMD's anymore roflmao"

{LOD}PITBULL · 03-15-2013, 16:14 | posts: 768 | Location: Upstate NY

Quote:

Originally Posted by RoadKillNL

Haswell has 1150 pins, x79 has 2011 pins.

LOL I was just going to say that...

Nono06 · **Memory**: 6GB (8-9-8-24-1T)@1870

Quote:

Originally Posted by Denial

I'm not talking about performance characteristics, I'm talking about processor errata. Intel documents it on the release of every stepping. Certain programming functions will literally be broken on certain processors. For instance Intel had a TLB bug in it's MMU on Core 2 which would cause hardlock ups of the processor when performing certain operations.

Here is an example of the documentation that Intel releases for programming around errata:

http://download.intel.com/design/mob...T/31407918.pdf

If you scroll down to page 40 you can see the list of all the problems and workarounds for them.

http://www.anandtech.com/show/2477/2

Anandtech did its own write up on the AMD version of the TLB errata.

Most of these problems aren't seen by end users because developers are aware and make changes in order to fix, or they are fixable via bios/microcode updates. This USB3 problem I guess is more significant and cannot be fixed via these methods but by a new stepping. But to pretend like processors never have problems and Intel is just now selling a faulty chip is nonsense.

+1. Bug free HW does not exist. As you said, most of the time SW workarounds (at different levels) correct those problems. On designs I'm working on, we are doing the same. If our low level drivers/BIOS cannot mask the problem we send "coding recommendations" to our customers to avoid any problem in the final products.

The USB 3 problem is not in the CPU but on the “Lynx Point” chipset. So maybe some MB manufacturers will propose a replacement when a fix will be available.

---TK--- · 03-15-2013, 16:21 | posts: 14,940 | Location: New Jersey, USA

Time to replace socket 1155, I'm buying

{LOD}PITBULL · 03-15-2013, 16:47 | posts: 768 | Location: Upstate NY

ok when is it saposed to come out. we all know that they are NOT in a rush to put new stuff out.I.E 2011 soc. NO update since chip wise.

-Tj- · *Last edited by -Tj-; 03-15-2013 at 17:00.*

Quote:

Originally Posted by {LOD}PITBULL

ok when is it saposed to come out. we all know that they are NOT in a rush to put new stuff out.I.E 2011 soc. NO update since chip wise.

27.May - 7.June

lol I already bought 2x8Gb Cl9 ddr3 2133mhz and Corsair H90 cooler, all I need is z87 and i7 4770k.. Hurry up intel