Wait for haswell "bugs" to be ironed out before buying?

[yayeti]

Is there any sense to wait to upgrade for any problems to be worked out or would any potential problem already have been fixed before release?

 

[inf64]

What bugs? Haswell works great. There is/was only one issue and that's USB 3.0 in combination with S3 standby and there are reports intel already fixed it( the bug was in the chipset).

 

[Meekers]

What bugs? Haswell works great. There is/was only one issue and that's USB 3.0 in combination with S3 standby and there are reports intel already fixed it( the bug was in the chipset).

The Haswells being sold now do not have the fix, they are still bugged. I for one am waiting until the new stepping before I consider getting a Haswell.

 

[smakme7757]

I would wait if there isn't a pressing reason to buy right now.

 

[Revolution 11]

The only reason I would wait for Haswell is to see if the i5-4670 and i7-4770 are the better choice due to TSX featuers instead of the K chips which have limited overclocking. I would also wait to see how the delidded K CPUs behave with regards to temperature, clockspeeds, and voltage.

Finally, if you don't have pre-Sandy Bridge CPUs, don't bother with Haswell.

 

[yayeti]

Wait til when, August? Right now I'm using a $400 Toshiba I bought from Best Buy 4 years ago.

 

[Ken g6]

The only reason I would wait for Haswell is to see if the i5-4670 and i7-4770 are the better choice due to TSX featuers instead of the K chips which have limited overclocking. I would also wait to see how the delidded K CPUs behave with regards to temperature, clockspeeds, and voltage.

I'm waiting for this, to figure out whether or not non-K CPUs allow the new BCLK straps, and also for more mobo choices.

If you're buying a pre-built system again, and you don't intend to overclock, there's no reason to wait.

 

[Revolution 11]

Well, I am talking about desktop CPUs. Haswell for mobile will be great, get a laptop with one as soon as you want.

Any desktop chip you get will be a great upgrade from that Toshiba. I suggest waiting for those who are torn between TSX features (no benchmarks yet) and overclocking. If you are going to delid, might as well wait for those results.

If you don't care about TSX, overclocking, or delidding, no reason to wait.

 

[Concillian]

Wait til when, August? Right now I'm using a $400 Toshiba I bought from Best Buy 4 years ago.

In this case, I wouldn't be waiting for anything.

 

[Concillian]

Well, I am talking about desktop CPUs. Haswell for mobile will be great, get a laptop with one as soon as you want.

Any desktop chip you get will be a great upgrade from that Toshiba. I suggest waiting for those who are torn between TSX features (no benchmarks yet) and overclocking. If you are going to delid, might as well wait for those results.

If you don't care about TSX, overclocking, or delidding, no reason to wait.

TSX is something you'll need to make a decision about. It's not something where the benefits are going to be realized in a month or two except perhaps for some critical business oriented software where time is money. This isn't something to wait for. It's something you have to decide about when you purchase.

De-lidding... we already know the IHS config is the same as Ivy Bridge and we can extrapolate the benefit to be similar. As a result we have a pretty good idea of the overclocking story.

On a budget air cooler it's going to be 4.0-4.3 depending on luck of the draw and a couple hundred more if you de-lid. In other words, a tad below Ivy that will be largely made up for by slightly better IPC.

People already have the info they need to make a decision.

 

[Edrick]

Some TSX numbers:

http://www.sisoftware.co.uk/?d=qa&f=ben_mem_hle&l=en&a=

 

[blackened23]

What bugs? Haswell works great. There is/was only one issue and that's USB 3.0 in combination with S3 standby and there are reports intel already fixed it( the bug was in the chipset).

I'll echo this sentiment. Haswell has no bugs, i'm really not sure what the OP is referring to.

 

[blackened23]

The Haswells being sold now do not have the fix, they are still bugged. I for one am waiting until the new stepping before I consider getting a Haswell.

You realize that the issue is completely trivial, right? It only affects certain USB sticks which aren't recognized after sleep - it doesn't cause data loss and merely requires a reboot. I guess it could be an issue if it affected USB 3.0 external HD's, but so far it seemingly doesn't.

 

[Edrick]

No bugs found yet on my system. And I have been putting it through the paces.

I can honestly say that going from this (3820 @ 4.1Ghz, 8GB 2133mhz, Corsair Pro SSD) to this (4770 @ 3.7Ghz, 8GB 2400mhz, Samsung 840 Pro), my system feels faster and does benchmark faster.

 

[Concillian]

You realize that the issue is completely trivial, right? It only affects certain USB sticks which aren't recognized after sleep - it doesn't cause data loss and merely requires a reboot. I guess it could be an issue if it affected USB 3.0 external HD's, but so far it seemingly doesn't.

It doesn't even require a re-boot.

It just means that if you have a USB drive connected and file open when you sleep, it won't initially be connected when it comes out of sleep, but will re-initialize and be available shortly after. This is only really an issue if you are launching applications from external drives connected via USB 3.0 AND use sleep.

 

[bronxzv]

Haswell has no bugs

good one, now have a look here: Haswell Specification Update

 

[blackened23]

good one, now have a look here: Haswell Specification Update

Well, let me correct myself since you're being quite literal about this. All CPUs have "bugs", but they are generally completely trivial in nature and transparent to the user. That is what I meant. There have been exceptions to this in the past, but Haswell doesn't have any issues which aren't transparent to the user. The USB issue was fixed, and that is pretty trivial in nature and is chipset related.

 

[blackened23]

It doesn't even require a re-boot.

It just means that if you have a USB drive connected and file open when you sleep, it won't initially be connected when it comes out of sleep, but will re-initialize and be available shortly after. This is only really an issue if you are launching applications from external drives connected via USB 3.0 AND use sleep.

Gotcha, thanks for the clarification. Either way I see it as a non-issue - as long as my external HDDs aren't affected, i'm good to go. Perhaps some think differently though - and if so, wait 2-3 weeks for the C1 revision of Z87.

 

[jcniest5]

Why is the 4770s only running at 3.1Ghz?

 

[Beavermatic]

The IvyBridge-E series processors will be out in September. They will be the 49xx series line, have upto 6cores/15MB L2 Cache. They'll be 10% faster than the Sandybridge-E series, and upto 20% against standard ivy bridge or haswell quad cores.

Of course, the Haswell-E series wont be out until late next year, so you have quite a bit of time to play around with the Ivybridge-E in between.


I think that would definitely be the better upgrade path than going to haswell, which from everything I've read and seen, was a letdown, especially for those interested currently on ivybridge or sandybridge.

 

[Meekers]

You realize that the issue is completely trivial, right? It only affects certain USB sticks which aren't recognized after sleep - it doesn't cause data loss and merely requires a reboot. I guess it could be an issue if it affected USB 3.0 external HD's, but so far it seemingly doesn't.

Why would I spend so much money to upgrade to a bugged cpu when my current cpu is not letting me down. The only reason I am even thinking of buying this joke of a release is because of chipset upgrades form x58. Especially coming from a c0 i7 920 I am hoping that the new stepping improves OCing to at least polish this turd a bit.

Pretty upset that I have waited this long for a new cpu and intel just gave us the finger.

 

[subject42]

What bugs? Haswell works great. There is/was only one issue and that's USB 3.0 in combination with S3 standby and there are reports intel already fixed it( the bug was in the chipset).

I just put together my Haswell system last night, and the first thing I tried to do was reproduce that issue with two USB3 thumb drives I had lying around. I wasn't able to reproduce the issue. Either I'm doing it wrong, or the problem only happens on a subset of drives, or Gigabyte (z87-ud3h) managed to work around the issue on their own.

 

[bronxzv]

Well, let me correct myself since you're being quite literal about this. All CPUs have "bugs", but they are generally completely trivial in nature and transparent to the user. That is what I meant. There have been exceptions to this in the past, but Haswell doesn't have any issues which aren't transparent to the user. The USB issue was fixed, and that is pretty trivial in nature and is chipset related.

it depends of the user, all the errata related to performance counters for example show well in VTune Amplifier, I endured it on pre-prod hardware and the issues are still not fixed

also all the errata related to gather are said to have a BIOS-based workarounds, I read it as more microcode than expected for gather which may well explain why gather is so incredibly slow

 

[oneofusjustin]

The IvyBridge-E series processors will be out in September. They will be the 49xx series line, have upto 6cores/15MB L2 Cache. They'll be 10% faster than the Sandybridge-E series, and upto 20% against standard ivy bridge or haswell quad cores.

Of course, the Haswell-E series wont be out until late next year, so you have quite a bit of time to play around with the Ivybridge-E in between.


I think that would definitely be the better upgrade path than going to haswell, which from everything I've read and seen, was a letdown, especially for those interested currently on ivybridge or sandybridge.

what socket will Ivy-E use? also will the be putting out a reasonably priced quad or is it just gonna be the $500 6 cores?

 

[WhoBeDaPlaya]

What bugs? Haswell works great. There is/was only one issue and that's USB 3.0 in combination with S3 standby and there are reports intel already fixed it( the bug was in the chipset).

Another issue is the incompatibility with 1st gen SandForce (1xxx) drives.