Upgrading 3930k to 4930k (HP Phoenix 800-050se)

Eggz

Distinguished
Hi All,

As the title suggests, I am interested in upgrading the processor in my HP ENVY Phoenix 800-050se by one generation, from an Intel i7-3930k to an Intel i7-4930k. This is because I can sell the 3930k and obtain a new 4930k for less than a $100 spread, and $100 is worth a 10% performance increase to me.

The issue is whether the newer Intel i7-4930k is compatible with my current motherboard. I am 100% certain it can be, but I cannot verify that it actually is. Newegg and various other venders warn upgraders going from Sandy Bridge-E to Ivy Bridge-E (i.e. from 3930k to 4930k) to verify whether their existing LGA 2011 motherboard will require a BIOS flash update in order for it to recognize the new processor. Because HP uses a mysterious motherboard in its high-end desktops, finding this information has thus far been impossible. I will outline what I have tried, and hopefully someone here will be able to suggest something, because I am out of ideas.

The motherboard in question is HP's "Pittsburgh2" (IPIWB-PB). Here is a link to the official HP specification sheet. Pittsburgh2 is in my computer (the Phoenix 800-050se), which is currently equipped with an Intel i7-3930k, and the Pittsburgh2 is also equipped in other HP computers utilizing the newer Intel i7-4930k (see e.g. Phoenix 810-150se). This is why I am 100% certain the motherboard can support both chips.

My (so far) unanswerable question is, "Are the BIOS the same on both machines?" If so, I will not need to update the BIOS, and the chip will work; otherwise, upgrading will require a BIOS update. HP's update page for the 800-050se lists no BIOS update. But neither that page nor the update page for the 810-150se lists the BIOS version installed on either computer. Effectively, I cannot lookup information about BIOS versions, which prevents me from comparing my BIOS version to that of another HP computer using a 4930k on the Pittsburgh2 motherboard.

My next step was calling HP for help, many times. All of the attempts failed. The first call was to technical support. A support agent said that he would not look up product information without a product's serial number, but the only way to obtain that number is to purchase a computer. This didn't help because I need the BIOS version for the 810-150se, but I own the 800-050se. Thinking I may have simply gotten a grumpy tech guy, I called again (and again, and again . . .) until I got the same answer from 5 different people. I repeated a similar phone call process with HP’s departments for customer service, sales, and parts. Each department referred me back to tech support, which refused to answer my question; namely, “Is the same BIOS version on both the 800-050se and the 810150se? If not, where do I get the update, because it isn’t on the HP website?”

The next step in my search lead me to research third-parties. I discovered that the Pittsburgh2 motherboard is a product of the Pegatron Corporation, a division of ASUS. So, I contacted Pegatron via e-mail - the only way for consumers to contact Pegatron. It turned out that Pegatron is an OEM company that by contract with HP cannot support motherboards used in HP products. They apologetically referred me back to HP, but the previous paragraph explained how that went. Then I contacted ASUS, but they told me the same thing as Pegatron. Finally, I called Newegg's sales people because I know they have experts to answer compatibility questions. Those experts, however, lack access to information on HP's Pittsburgh2 motherboard.

Apparently, the answer lies with HP but HP lies about having it. I have never had such a difficult time finding such basic information. All motherboards I have owned in the past have a simple webpage where this information is easily accessible. If anyone has any information that will answer my question, I would greatly appreciate it!

Without the information, my only option will be to spend nearly $600 on an Intel i7-4930k to test compatibility. But if it the test fails, then I will have to eat money on returning a used processor (the 4930k), or else spend an additional $300 on a normal ASUS motherboard. But that would set me back more than $1,000 after tax, which is 10-times more than I should spend on this upgrade. Clearly, this simple but illusive piece of information would be extremely helpful for me. Please help!

Thank you

-Eggz
 

Eggz

Distinguished


Because of my insatiable desire for speed! I also have the money and enjoy tinkering :) . . . But in slightly more detail, it's mainly because of the benefit every bit of speed gives when processing hundreds, and sometimes thousands, of RAW photo images. Even though I have lots of RAM, large SSDs, high end video cards, etc., such large-scale processing invariably takes a long time. There's really no way around that. So for me, it's worth it for stuff like this, especially since there's still a solid market of buyers on eBay for my used 3930k.
 

Specops125

Distinguished
Dec 17, 2013
209
0
18,760
If the computer sitting in your room existed before Ivy Bridge-E existed/had its tech specs and data released, I would imagine that your mobo would need the BIOS update. After all, this was the case with Asus's Asus branded LGA 2011 motherboards. You can get a rough idea of when it was manufactured by looking at the oldest installed software by HP in Control Panel.
 

Eggz

Distinguished


Yeah, that very suspicion lead me on this wild goose chase to begin with. I've started a thread on HP's forum to see if anyone with a i7-49xx on a Pittsburgh2 motherboard will post their BIOS for me to check against the latest HP BIOS flash update for that motherboard.

If anyone replies to confirm that they have the same BIOS version as the available BIOS update through HP, I'll be good to upgrade; otherwise, I'll just hold out until Haswell-E. Waiting's beginning to make more and more sense. I'm beginning to think that "fast" doesn't really exist when it comes to photo editing that involves MANY high resolution images. The more accurate description is "slowness minimization." I have a very capable system, and it still involves a lot of waiting. That is despite having an Intel i7-3930k, 32 GB of RAM at 2133 Mhz (which does fill up during editing sessions), a 750 GB Samsung EVO, a 2TB RAID 1 array for media, a GTX 770, yada yada. Involved editing processes are simply very complicated and resource hungry - and that's that. Going 3930k --> Haswell-E, with its DDR4, seems to offer a more significant reduction in slowness than 3930k --> 4930k.

That being said, I would still upgrade to a 4940k if the BIOS flash will enable doing so. Eating $100 net cost on the upgrade is just a drop in the bucket compared to what I've invested in the system, and a 10% increase in performance remains significant. If you think about it, 10% of a 3930k is substantial.

Also, Sandy-E's main weakness is single-core performance. The Ivy-E processors have much better single core performance, and that is primarily where photo editing gets hung up. Adobe software does a lot of amazing things that can't be delegated very well to multiple threads, so the work often gets dumped on a single core. I very frequently see 14% usage because one core is maxed while the others idle.

My research even shows that the lower-grade i7-4770k outperforms the 3930k's single core performance. When I first found this out, I was shocked because a corollary was that upgrading to a Xeon would actually decrease performance for my applications because most Xeons have more slower cores. Ivy-E, on the other hand, seems to be the best option for a processor that has more than 4 cores, maintains high single core performance, and manages to be affordable to individuals (as opposed to institutions). Sad face if I can't upgrade.
 

Specops125

Distinguished
Dec 17, 2013
209
0
18,760
It will be cool to see what the 8 core Haswell-Es clock speed/single core performance will be. If the 5930K and 5960X will be 8-core, and if the 5820K is made 6-core, I suppose the entry-level E processor will take a bigger market share on gaming rigs as compared to SE-E and IB-E. It's interesting that you say that about Adobe, I'm guessing you are talking about CS6? I thought that was usually held up as a prime example of software that can make use of several cores. I suppose either way that puts you in a good position for Haswell-E: The higher clocked 6-core CPU would save you a lot of money, and if the 8-core would be better for you, you will now have that option from Intel for the first time on their consumer CPUs.
 

Eggz

Distinguished


Agreed!

Not to derail my own thread, and readers interested exclusively in the title's topic should skip this post --

I don't mean to paint the wrong picture about Adobe. Their software (primarily Lightroom and Photoshop in my case) can utilize all 12 of my logical processors. But much of what I do simply doesn't.

From researching the Adobe's processor utilization capabilities, there are apparently certain tasks that actually slow down or even fail when delegated via multiple threads. This is because multithreading is sometimes less efficient, and other times it's just impossible.

I found in my searches two particular analogies to be quite helpful - one for each situation. The first pertains to situations where delegation turns out to be impossible: "Just because one woman can produce a baby in nine months doesn't mean that nine women can produce a baby in one month." The second pertains to situations where delegation turns out to be inefficient: "Just because one person can solve a rubix cube in 10 minutes doesn't mean 10 people can solve it in one minute. In fact, assigning a single puzzle to 10 people will likely slow things down considerably." Keep in mind that my level of understanding is not vast or extensive when it comes to scripting C languages to efficiently multithread applications, but the concept seems intuitive to me.

In my experience, the things in Photoshop that resemble the "rubix cube" are the content-aware features. The computer samples an image and makes a best guess at what would fill in a selected portion of the image. Seeing it happen is actually quite fascinating! Another related feature I use a lot is the panorama creation tool, which seems to use content-aware technology to artfully cut multiple photos and seamlessly stitch them into one giant image file. Each of these tasks alone can utilize nearly all of my 32 GB of RAM and take a while to complete.

During processing, only one of my 12 logical processors runs (100%) while the other 11 idle, which is specific only to certain tasks like the two I mentioned. Many other features in Photoshop or in Lightroom distribute tasks more evenly across the processor - and in some cases - to the GPU. But because tasks like panorama and content-aware fill account for a disproportionate amount of my processing time, obtaining best in class single-core performance could represent an attractive performance gain.