USB 3.0 problems for Intel's Haswell

[Fjodor2001]

Only storage devices really. And HDs was just the extreme example. Since they are terrible even today.

Only storage USB devices resume from S3 sleep? What proof do you have of that?

 

[jcniest5]

Major recall again! This time, it's the CPU. I will wait until a new stepping to jump on the bandwagon.

Edit: Oh, chipset, too? Darn...bad!!

 

[ShintaiDK]

Only storage USB devices resume from S3 sleep? What proof do you have of that?

You do know the entire thread is only about mass storage devices connected to USB3, right?

 

[Fjodor2001]

You do know the entire thread is only about mass storage devices connected to USB3, right?

The thread is about USB 3.0 problems on Haswell (which later was confirmed to actually be present on the related Lynx Point chipset instead). The exact details are still being uncovered as more information about the problem is discovered.

What I question is whether is makes sense to assume that only USB mass storage devices are affected by this problem, since other USB devices resume from sleep too. Do we know enough about the problem yet do definitely say that other types of USB devices are not affected too? In that case what technical aspect of the problem proves that is the case?

If you look at the statement in the original post it says:

"In the document, Intel purportedly informs manufacturers that a system with Haswell inside, when waking from S3 sleep mode, will experience issues with devices connected through USB 3.0."

So couldn't it be that USB mass storage devices are just one type of USB devices that are affected by the S3 resume problem?

 

[ShintaiDK]

Uhm...from your own post with more details.

The problem in question effects some USB flash drives, and involves the chipset ceasing to poll some drives when a Z87 system enters an S3 sleep state. This means that the USB drive does not wake back up with the system, and you need to reconnect the drive in order to continue working with it.

 

[Fjodor2001]

Uhm...from your own post with more details.

But how do you know that isn't just one example of the USB resume problem?

What technical aspect of the problem makes it likely that only USB mass storage devices should be affected? That's what I wonder.

 

[NTMBK]

is it just me..or does anyone else not care about this usb bug....

i have never in my 24 years ever used sleep mode

Working in the library on your laptop on your essay on a USB pen- then oh crap, you need to run to lecture! Shut lid on laptop, dash over to lecture theatre, wake laptop from sleep, and boom, no essay.

 

[Magic Carpet]

^. Employ better habits, i.e. use portable software with auto-save feature. Helps with crashes, power surges too

 

[NTMBK]

^. Employ better habits, i.e. use portable software with auto-save feature. Helps with crashes, power surges too

Don't worry, I'm not saying that it's the most sensible way to use a laptop. All I'm saying is that there is a lot of people who will do this, and would be very pissed off if they lost their work.

 

[phillyman36]

So is it the cpu or the mobo that this will get fixed in the next stepping in July ? Someone said its the cpu some said its the mobo.

 

[ShintaiDK]

So is it the cpu or the mobo that this will get fixed in the next stepping in July ? Someone said its the cpu some said its the mobo.

Mobo.

Its the chipset, not the CPU that contains the bug.

 

[podspi]

^. Employ better habits, i.e. use portable software with auto-save feature. Helps with crashes, power surges too

Reminds me of when Steve Jobs told people they were holding the iPhone wrong

This is a major issue that probably nobody will be affected by because OEMs will probably just add additional controllers to make up for the problem.

IMHO, this is an effect of the lack of competition in the CPU market. Intel knows they can piss off their partners (forcing them to eat the bug), what are they going to do, switch to ARM?

 

[ShintaiDK]

IMHO, this is an effect of the lack of competition in the CPU market. Intel knows they can piss off their partners (forcing them to eat the bug), what are they going to do, switch to ARM?

BS.

Also lookup past bugs from all vendors. You would be surprised and horrified how many there are and what they cover.

 

[Sherlockwing]

Reminds me of when Steve Jobs told people they were holding the iPhone wrong

This is a major issue that probably nobody will be affected by because OEMs will probably just add additional controllers to make up for the problem.

IMHO, this is an effect of the lack of competition in the CPU market. Intel knows they can piss off their partners (forcing them to eat the bug), what are they going to do, switch to ARM?

Simpler/Cheaper than that, Tweak the Bios to use S2 instead of S3 as the default sleep state. Plus from the leaks so far the early Haswell Mobiles are the Quad core HT parts with 47/57W TDP that go in Gaming Laptops/Desktop Replacements so Battery life won't be the no.1 concern of those users.

 

[386user]

i also believe i read that ultralow voltage processors are coming a month afterwards...but they have already been in production since q4 last year....so, who knows

 

[Edgemeal]

BS.

Also lookup past bugs from all vendors. You would be surprised and horrified how many there are and what they cover.

Does it really matter who's had bugs/faults in the past or even currently?
a bug is A BUG! D:

Even if the bug doesn't effect me today how do I know it wont effect me in the future, or when I try to use some other hardware that hasn't even been released yet? (I've seen this happen before!)

Buggy hardware is also harder to re-sell and fetches a lower price.

If this bug was from anyone other then Intel your attitude (and many others here) would be totally different, but ohhh its our beloved "Intel" so it's no big deal, LOL!

Why can't Intel just fix this before they release it?

 

[Idontcare]

Reminds me of when Steve Jobs told people they were holding the iPhone wrong

Ironically I had drafted a post expressing the same but then I got distracted and forgot to submit it

IMO this is one of those cases where the proof of the pudding is in the eating...meaning if this usage-rate of this feature (actively editing files on USB drives from computers that are also prone to entering the sleep state) were truly in the realm of the "negligible and irrelevant few" then the feature itself would be deprecated and unsupported such that the said "bug" no longer warrants investing further resources to mitigate.

Instead, the presence of the "bug" is downplayed by way of downplaying the relevance of the feature that the bug undermines. "You just need to hold your phone differently" "You shouldn't be leaving unsaved work on a computer that might go to sleep" etc.

The justification of the feature then becomes one of justifying eliminating the bug rather than one of justifying enabling the feature in the first place.

"Oh well, the consumer didn't really need that feature anyway..."

 

[podspi]

BS.

Also lookup past bugs from all vendors. You would be surprised and horrified how many there are and what they cover.

I'm not talking about the bug, I'm talking about how Intel can force the OEMs to take shipment of defective product. If you honestly think competitive pressures do not affect Intel, you're just wrong. There are many dimensions of competition, and negotiating position of partners is one of them.

Simpler/Cheaper than that, Tweak the Bios to use S2 instead of S3 as the default sleep state. Plus from the leaks so far the early Haswell Mobiles are the Quad core HT parts with 47/57W TDP that go in Gaming Laptops/Desktop Replacements so Battery life won't be the no.1 concern of those users.

Actually, couldn't they just default to S0ix? My stupid ATIV 500t crashes every time it comes out of S3 -- so I just disabled sleep entirely and I still get 10hrs+ battery life.

 

[vampirr]

http://www.fudzilla.com/home/item/30720-haswell-usb-30-issue-requires-hardware-fix

 

[ShintaiDK]

I'm not talking about the bug, I'm talking about how Intel can force the OEMs to take shipment of defective product. If you honestly think competitive pressures do not affect Intel, you're just wrong. There are many dimensions of competition, and negotiating position of partners is one of them.

How many TLB Phenoms got recalled? How many 1.13Ghz P3s? How many nVidia mobile GPUs? How many ATI and nVidia chipsets?

The list is endless. This is how the industry works.

Just look at software, they got the biggest get out of jail freecard on this planet.

 

[Pheesh]

podspi: the only thing 'forcing' OEM's to take a product that has any bugs is THEIR competitors, not Intel. They all want to be first to market, if they can already mitigate the bug they are going to take that market timing advantage every single time. If they don't, their competitor will.

 

[Obsoleet]

Don't worry, I'm not saying that it's the most sensible way to use a laptop. All I'm saying is that there is a lot of people who will do this, and would be very pissed off if they lost their work.

Why isn't it sensible? We're to change usability based on buggy hardware that none of my old computers exhibits? Blame the user? Screw that. There's always AMD. No reason to put up with this. It'll be interesting if Apple ships Macbooks with Intel's f****up embedded.

 

[frozentundra123456]

I'm not talking about the bug, I'm talking about how Intel can force the OEMs to take shipment of defective product. If you honestly think competitive pressures do not affect Intel, you're just wrong. There are many dimensions of competition, and negotiating position of partners is one of them.



Actually, couldn't they just default to S0ix? My stupid ATIV 500t crashes every time it comes out of S3 -- so I just disabled sleep entirely and I still get 10hrs+ battery life.

I think it must be more a matter of being caught by surprise. Do you think Intel would really want the bad press for this if they could have done anything to prevent it? Somebody screwed up in testing/validation. I actually think it might have been better for Intel to just eat the cost and delay introduction till the next stepping. Personally, I never let my laptop go to sleep, I save and shut down every time I leave it, so the problem would not really affect me. But I am sure some users would be affected, and where it seems the biggest problem is in tablets/laptops, which is exactly where Haswell is supposed to excel.

 

[ShintaiDK]

Remember the first batch is desktop CPUs. Might never be a laptop with this bug.

 

[cytg111]

Why isn't it sensible? We're to change usability based on buggy hardware that none of my old computers exhibits? Blame the user? Screw that. There's always AMD. No reason to put up with this. It'll be interesting if Apple ships Macbooks with Intel's f****up embedded.

Yeaa, see, even though the bug is a BUG, if I personally had to choose between a haswell rig WITH the bug and an amd FX-8* ? It would be buggy haswell ....