Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
Anime Schoolgirl
Nov 28, 2002

VulgarandStupid posted:

You or your friend should be buying up 5775c's, just saying.
pretty much this, unless you like the idea of spending another broadwell-c's worth of money for kaby/cannonlake-c

which apparently your friend does, so w/e

Adbot
ADBOT LOVES YOU

Potato Salad
Oct 23, 2014

nobody cares


I learned today he is actually trying to throw money down a hole. This is his express purpose. We're going car shopping later :homebrew:

Anime Schoolgirl
Nov 28, 2002

is it a windfall or debt fraud

Krailor
Nov 2, 2001
I'm only pretending to care
Taco Defender
Well hey, if that's the case you can feel free to send a 5775c my way...

Marinmo
Jan 23, 2005

Prisoner #95H522 Augustus Hill

Potato Salad posted:

I learned today he is actually trying to throw money down a hole. This is his express purpose. We're going car shopping later :homebrew:
He's more than welcome to contact me to get rid of his 4790k so he can buy a ricier car. I'm not saying he's dumb, he probably got his reasons, but I am saying I'd love to rip that thing off him for cheap.

redeyes
Sep 14, 2002

by Fluffdaddy

Don Lapre posted:

My kids elementary school only uses chromebooks and google docs.

My kids school uses mostly chromebooks and google docs, and some PCs for advanced stuffs. So yeah, chromebooks bitches. Google is killing it.

AND I happen to know that at least half the schools in my town use them as well.

Lets be honest here. The chromebook is the best device for a learning environment ever. Seriously, low cost, no viruses, functional. Perfect.

redeyes fucked around with this message at 01:57 on Oct 1, 2015

Star War Sex Parrot
Oct 2, 2003

Holy poo poo make a thread to argue about Chromebooks

Backyarr
Jun 6, 2006
There's a pirate in your backyard!

Fallen Rib
Speaking of desktop Broadwells, I was thinking about putting together a new HTPC since my current Atom 330 + Nvidia Ion setup is getting a bit old. I've been looking at the i5-5675C as the ideal candidate with it's integrated Iris graphics and as a plus, all compatible mini-ITX Z97 motherboards have a full-length PCI-E slot for any eventual future GPU upgrades.

Everything was already in a shopping cart waiting for my credit card details, but then I saw this:
http://news.softpedia.com/news/the-intel-broadwells-have-a-major-design-defect-when-running-office-2016-493094.shtml

So apparently the desktop Broadwell family has a design flaw of some sort. Anyone know anything more about this, or have any experience with the Broadwells to share?

VulgarandStupid
Aug 5, 2003
I AM, AND ALWAYS WILL BE, UNFUCKABLE AND A TOTAL DISAPPOINTMENT TO EVERYONE. DAE WANNA CUM PLAY WITH ME!?




Backyarr posted:

Speaking of desktop Broadwells, I was thinking about putting together a new HTPC since my current Atom 330 + Nvidia Ion setup is getting a bit old. I've been looking at the i5-5675C as the ideal candidate with it's integrated Iris graphics and as a plus, all compatible mini-ITX Z97 motherboards have a full-length PCI-E slot for any eventual future GPU upgrades.

Everything was already in a shopping cart waiting for my credit card details, but then I saw this:
http://news.softpedia.com/news/the-intel-broadwells-have-a-major-design-defect-when-running-office-2016-493094.shtml

So apparently the desktop Broadwell family has a design flaw of some sort. Anyone know anything more about this, or have any experience with the Broadwells to share?

Seems like an awful lot of power for an HTPC, but more concerning than that is if you're actually going to run Office on your HTPC.

Backyarr
Jun 6, 2006
There's a pirate in your backyard!

Fallen Rib

VulgarandStupid posted:

Seems like an awful lot of power for an HTPC, but more concerning than that is if you're actually going to run Office on your HTPC.

No offense bro, but read the article:

quote:

The bug appears to users running on Windows 7, 8.1, 10, multiple Linux versions and even crashing a machine that was running a Linux virtual box

Somehow, having a Linux setup crash all the time doesn't sound appealing to me since it defeats the purpose of a Linux install. I was also planning on having it eventually run a web server and a home file server, I probably don't need all that power but I kinda want to have it last for at least 6-7 years.

With some Googling, I came up with these:
https://www.reddit.com/r/hardware/comments/3meznc/design_defect_in_i55675ci75775ci75700hq/
https://bugzilla.kernel.org/show_bug.cgi?id=103351
https://communities.intel.com/thread/80822

Incredulous Dylan
Oct 22, 2004

Fun Shoe
Received some new parts yesterday to update my build. 6700k and ASUS Maximus Ranger mobo. Was impressed with how well the chip overclocks - just running the ASUS auto over clock had me at a stable 4.8 GHz. For anyone with that mobo I just want to note that specifically the new BIOS that came out yesterday, 0905, had me unable to boot (Q Code 99). I flashbacked to the BIOS before that and everything was solved.

BobHoward
Feb 13, 2012

The only thing white people deserve is a bullet to their empty skull

Did you actually read this one all the way through? I don't think you did, because you still seem to be in Chicken Little mode. One of the commenters on that bug with a MSI laptop got directed to this:

https://forum-en.msi.com/index.php?topic=261054.msg1498718#msg1498718

and applying the BIOS update fixed their problems. Someone else who didn't have a MSI laptop extracted the CPU microcode from the MSI UEFI update and managed to use it to fix their problems too.

So: brand new Intel CPU needs a microcode update to fix an issue. This is not actually unusual, it's rather common. Microcode updates are typically delivered via BIOS/UEFI updates but can be loaded by the Linux kernel too.

Backyarr
Jun 6, 2006
There's a pirate in your backyard!

Fallen Rib

BobHoward posted:

CPU needs a microcode update
Well, that's great! But I seriously hope someone puts out a, y'know, *official* microcode update. Thanks man, seems like the kernel.org thread updated quite a bit since I read it last.

Anime Schoolgirl
Nov 28, 2002

The bios updates are official, though. As in Intel gave them the microcode bug fixes and all.

DrDork
Dec 29, 2003
commanding officer of the Army of Dorkness
Sounds like it just hasn't made its way out through all the mobo manufacturers, though. Which, again, isn't anything new--some companies are a lot more prompt about updating and releasing BIOS versions, while others slow roll it.

JawnV6
Jul 4, 2004

So hot ...

Backyarr posted:

But I seriously hope someone puts out a, y'know, *official* microcode update.
You think Asus is signing them?

Backyarr
Jun 6, 2006
There's a pirate in your backyard!

Fallen Rib

JawnV6 posted:

You think Asus is signing them?

I was actually thinking more along the lines of "all manufacturers that claim to support Broadwell should publish BIOS updates" so as not to rely on a workaround using microcode salvaged from a single BIOS update.
I'm really glad a solution is possible, though.

nzspambot
Mar 26, 2010

Is anyone running 32Gb on Z170H?

Under Windows 10 on a Asus Z170H-Deluxe it was well unstable as gently caress when trying to upgrade Windows (Insider Builds)

Finger could be pointed at Nvidia (Running a GTX-970)

4xKingston something at 2666 btw

Very stable with 16Gb also

Jan
Feb 27, 2008

The disruptive powers of excessive national fecundity may have played a greater part in bursting the bonds of convention than either the power of ideas or the errors of autocracy.

nzspambot posted:

Is anyone running 32Gb on Z170H?

Under Windows 10 on a Asus Z170H-Deluxe it was well unstable as gently caress when trying to upgrade Windows (Insider Builds)

Finger could be pointed at Nvidia (Running a GTX-970)

4xKingston something at 2666 btw

Very stable with 16Gb also

Bad RAM? Have you tried all 4 modules at 16gb?

Don Lapre
Mar 28, 2001

If you're having problems you're either holding the phone wrong or you have tiny girl hands.

nzspambot posted:

Is anyone running 32Gb on Z170H?

Under Windows 10 on a Asus Z170H-Deluxe it was well unstable as gently caress when trying to upgrade Windows (Insider Builds)

Finger could be pointed at Nvidia (Running a GTX-970)

4xKingston something at 2666 btw

Very stable with 16Gb also

Turn off xmp

LiquidRain
May 21, 2007

Watch the madness!

My first thought is that he's running 1T and not 2T. Running 1T with 4 modules is asking for trouble.

We have a subforum for these kinds of things, but this isn't the first (nor will it be) the last.

Anime Schoolgirl
Nov 28, 2002

Even most 2 module XMP setups default to 2T.

The speed difference is insignificant in the age of DDR3/DDR4. Always have 2T enabled.

redeyes
Sep 14, 2002

by Fluffdaddy

nzspambot posted:

Is anyone running 32Gb on Z170H?

Under Windows 10 on a Asus Z170H-Deluxe it was well unstable as gently caress when trying to upgrade Windows (Insider Builds)

Finger could be pointed at Nvidia (Running a GTX-970)

4xKingston something at 2666 btw

Very stable with 16Gb also

I have 10 with an Asus z170-A and 32GB Cricual DDR4 PC2133. I booted and ran at stock speed and voltage and have been stable now for more than a week. I'd say its the memory speed or memory itself is bad.

Jan
Feb 27, 2008

The disruptive powers of excessive national fecundity may have played a greater part in bursting the bonds of convention than either the power of ideas or the errors of autocracy.

Anime Schoolgirl posted:

Even most 2 module XMP setups default to 2T.

The speed difference is insignificant in the age of DDR3/DDR4. Always have 2T enabled.

Eh, in the rare scenarios where RAM timing still actually matters, command rate is the only RAM timing that honestly makes a difference, with CAS latency being so high now.

Of course, those scenarios don't actually exist anymore outside of synthetic benchmarks so that's kind of moot. But still, given the choice of lower CAS or lower CMD, I'd pick the latter -- given the choice of lower CAS/CMD or more actual memory, always pick the latter.

Rastor
Jun 2, 2001

ARM vs. Intel:
Qualcomm enters server CPU market with 24-core ARM chip

DrDork
Dec 29, 2003
commanding officer of the Army of Dorkness
Wow, that chip is massive.

Combat Pretzel
Jun 23, 2004

No, seriously... what kurds?!
So, is it actually faster than an Intel chip with equivalent die surface or power consumption? Or is this just a novelty crossed with a crowbar?

champagne posting
Apr 5, 2006

YOU ARE A BRAIN
IN A BUNKER

"The limiting factor is how much power you can get in the building," Gavrielov of Xilinx said.

Limit not actual, power that can enter a building is much higher than datacenter use.

JawnV6
Jul 4, 2004

So hot ...

DrDork posted:

Wow, that chip is massive.
Giant pinout, but the power delivery network looks pretty limited. About the die size of a comparable Xeon?

Potato Salad
Oct 23, 2014

nobody cares


More competition in the compute market would be awesome, but God help me if I have to start supporting multiple architectures or start staving off devs who want to build their cloud stack on ARM.

Durinia
Sep 26, 2014

The Mad Computer Scientist

Combat Pretzel posted:

So, is it actually faster than an Intel chip with equivalent die surface or power consumption? Or is this just a novelty crossed with a crowbar?

You want actual information from a marketing roll-out? :ohyou:

Serious comment: faster by what measurement? Part of the promise of ARM is chips that do better on specific things that Intel is not great at, but maybe not have quite the breadth of Xeon.

The first design is a test platform - there's quite a ways to go with these guys - but QCOM has shown they can beat Intel in other markets. Of course, server is now home turf for x86...

BobHoward
Feb 13, 2012

The only thing white people deserve is a bullet to their empty skull
(This is not my original observation.) It's possible that Qualcomm is trying to target that chip at niches where compute power isn't important, but lots of IO is. That's a more achievable target than trying to compete straight up with Intel's CPU cores (*), and it gives them some room to undercut on price. If you were to buy a Xeon system with as much IO as that package implies, it's 2-socket for sure and possibly 4-socket or more. As they've grown more and more dominant, Intel's premium for 2+ socket enabled E5/E7 CPUs has skyrocketed, particularly 4 socket and above. If there are enough customers interested in a 1-socket IO monster at a bargain price, that's not a bad way to get a foothold. Exploiting a niche the monopolist isn't serving well is the way to go.

(*) - the difficulty of trying to compete with Intel's IO and interconnect should not be underestimated either, but QC is avoiding some of that by keeping the interconnect fabric on-die. Depending on delivering the equivalent of QPI in the gen 1 product would not have been smart.

Kerbtree
Sep 8, 2008

BAD FALCON!
LAZY!
So, upgrade question:
I've currently got an i3-2100, I can get an i5-2500t rather cheap. would I see a noticeable improvement in doing a drop-in replacement, or should I bank the cash to hasten a full board & chip swap further down the line?

Combat Pretzel
Jun 23, 2004

No, seriously... what kurds?!

Durinia posted:

Serious comment: faster by what measurement?
Uh, performance? I suppose in this particular case, web requests per second and whatever database metric is important?

Point is, if I have this fancy 24 core ARM CPU that eats 120W, and it gets blown to smithereens by the 8 core Xeon at 120W, what good is the ARM, except likely being cheaper in acquisition (if at all).

Durinia
Sep 26, 2014

The Mad Computer Scientist

Combat Pretzel posted:

Uh, performance? I suppose in this particular case, web requests per second and whatever database metric is important?

Point is, if I have this fancy 24 core ARM CPU that eats 120W, and it gets blown to smithereens by the 8 core Xeon at 120W, what good is the ARM, except likely being cheaper in acquisition (if at all).

Yeah, that's pretty much my point. The optimal configuration is very different, depending on workload. If it's thread/memory request bound, then lots of medium cores likely outperform a few big cores. If it's memory bound, more DDR channels, or it could be IO (like BobHoward said above). The goal is not to beat Intel at everything, just beat them at one relatively lucrative thing. Let some other vendor take aim at the other markets - there's plenty to go around.

BobHoward posted:

Exploiting a niche the monopolist isn't serving well is the way to go.

Explained better than I've managed to do it, apparently.

mobby_6kl
Aug 9, 2009

Well do they beat Intel at anything that people actually care about? I dunno, but I wouldn't automatically assume that.

Rastor
Jun 2, 2001

I don't think even they know fully yet:

quote:

The company is shipping the part to big customers now for testing, he said, though he declined to name them. He also wouldn't say when a product will be available commercially; Qualcomm will provide an update on that during the next year,

Still, it's been working on the part for two years and demonstrated it Thursday running a version of Linux, with the KVM hypervisor, streaming HD video to a PC. The chip was running the LAMP stack -- Linux, the Apache Web server, MySQL, and PHP -- and OpenStack cloud software.

Chandrasekher was joined by the CEOs of Mellanox and Xilinx, who are working with it to build a complete server platform. Mellanox is designing network cards to work with the SOC, while Xilinx said it will build programmable chips to speed up particular workloads.
My guess is that their angle is superior performance per watt in heavily multithreaded applications.

It sounds like at least some of the big names in cloud computing you would expect are taking a look, they are very experienced with spreading loads to make use of multiple cores.

Skandranon
Sep 6, 2008
fucking stupid, dont listen to me
They also don't need to soundly defeat Intel in any specific performance metric, as they can also seriously compete on cost/core.

champagne posting
Apr 5, 2006

YOU ARE A BRAIN
IN A BUNKER

Skandranon posted:

They also don't need to soundly defeat Intel in any specific performance metric, as they can also seriously compete on cost/core.

This is not a metric you even bother looking at when designing servers that run 24/7. Perfromance per watt is still king and this is unlikely to change.

Adbot
ADBOT LOVES YOU

canyoneer
Sep 13, 2005


I only have canyoneyes for you

Kerbtree posted:

So, upgrade question:
I've currently got an i3-2100, I can get an i5-2500t rather cheap. would I see a noticeable improvement in doing a drop-in replacement, or should I bank the cash to hasten a full board & chip swap further down the line?

Unless "cheap" means "free", don't do it.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply