[START TASK]

[Commence writing 980X Gulftown review] [#6876584] - generate specifications - copy/paste marketing slides - set stock frequencies, run benchmark - overclock slightly, run benchmark - mention price - conclusion: best performance, highest price.

[END TASK]

[START TAST]

[Commence writing Gulftown review] [#6876584] - ...

...

Surely, laying my hands on the latest CPU architecture, knowing it comes at a retail price of €1000 and it's in fact just the best a (desktop)-man can get, is exciting, but sometimes I forget that I'm not the only one testing this product. No, there are thousands of people all around the globe having the exact same feeling ... and willing to do the exact same type of article. The question is: does it matter whether I produce yet another run-of-the-mill Gulftown review or can I provide our readers with some information that other reviews don’t have and present it in a creative manner. Well, in this case I decided to go for the latter.

However, before we continue this madness, allow me to go over the Gulftown basics just one more time. For those who crave for performance, just like me, this new Intel product is interesting for two simple reasons:

It has SIX cores and hyper threading

build with 32nm process, not 45nm

Judging from the overclocking results obtained with the Intel Clarkdale architecture, which was the first 32nm variant of the Intel Core series, that last argument basically comes down to: better overclockability than what we were used to from the Bloomfield samples. I'm sure everyone knows what I am talking about here: the 45nm Bloomfield series hardly scaled with lower temperatures and hardly scaled with increased voltage and with those two being the main tools of an extreme overclocker, I suppose you can already see that these CPUs were not an overclocker's dream. Luck of the draw mainly, although obtaining a high-end mainboard did help to get slightly higher results.

The first argument needs no explanation: the more cores we have, the better we can see multi-threaded benchmarks, such as 3DMark Vantage or Wprime, scale.

Test setup

Massman's Core i7 980X Review System
CPU	Intel Core i7 950 Intel Core i7 960 Intel Core i7 980X
Cooling	Nventic Mach2 phase-change Kingpincooling Dragon F1 EE OCZ Freeze cooling paste
Mainboard	Gigabyte X58A-UD7
Videocard	MSI GTX275 Lightning MSI Radeon HD5870 Lightning
Memory	3x2GB Corsair Dominator 1600MHz CL8-8-8-24 1.65V, v3.1 (p/n: CMD8GX3M4A1600C8)
Storage	Western Digital WD3200KS Acard ANS-9010
PSU	Enermax Revolution 1050W

Test Methodology :

To test the performance, we used the following benchmarks.

Cinebench 10: Single and Multi CPU

Hexus Pifast

Resident Evil 5 DX9/DX10: 1280x720x32bit, no quality

SuperPi 1M and 32M

TechArp X264

Winrar

Wprime 32M and 1024M

3DMark03

3DMark05

3DMark06

3DMark Vantage

Aquamark3

Operating System: Windows XP 32-bit, Windows Vista 32-bit and Windows 7 64-bit. (Depending on benchmark application - same OS for same benchmark for each platform)

Intel Core i7 960/980X Setings


(Click to enlarge)

The GIGABYTE GA-X58A-UD7

A few weeks ago we received Gigabyte's most high-end X58-based mainboard, the X58A-UD7. This board offers the features that come with the 333 technology, tested here before at Madshrimps. What is interesting about this board is that Gigabyte has implemented all the tweaks and tricks they've learned in the past year of Bloomfield overclocking on X58. This means that, for instance, voltage modifications have been implemented in the PCB directly. To give a more concrete example: the voltage modification necessary to improve the stability of the PCIe lane during QPI overclocking (or: high BCLK frequency) used to be an end-user modification, but in this UD7 it's an option in the bios! In theory, this means that we'll be able to push far beyond 222MHz BCLK, which certainly comes in handy when testing locked-multiplier processors.

Before we gave this board a spin using the new 980X Gulftown, we had a quick run testing the Core i7 950 and the Core i7 960, two of Intel's latest Bloomfield products. In addition, both have locked multipliers, so the theoretical 222MHz+ will certainly come in handy.

Intel Core i7 950 Results

Single Stage: -25°C

3DMark03 at 210x23 = 4807MHz (HT on): 105895

3DMark05 at 210x23 = 4807MHz (HT on): 36934

3DMark Vantage at 210x23 = 4807MHz (HT on): P25842

PCMark05 at 219x23 = 5037MHz (HT off): 28415

On air cooling, I was only able to reach 220MHz BCLK, even with elevated PCIe frequency and IOH/VTT voltage levels. From the moment the temperatures went below zero, it was fairly easy to increase the BCLK frequency to 235MHz stable. This particular Core i7 950 sample proved to be quite a decent clocker even on phase-change hitting 4.8Ghz stable with all cores enabled.



Liquid Nitrogen: -100°C

3DMark05 at 228x23 = 5244MHz (HT on): 42643

3DMark06 at 228x23 = 5244MHz (HT on): 30785

Aquamark3 at 230x23 = 5290MHz (HT on): 372540

Wprime 32M at 235x23 = 5407MHz (HT on): 4.405 seconds

Wprime 1024M at 234x23 = 5370MHz (HT on): 2 minutes 20.421 seconds

On liquid nitrogen we noticed the same BCLK was at around 238MHz, making 235MHz fully stable. This processor was, all in all, quite a good sample reaching almost 5.4GHz Wprime stable. Due to the coldbug problem and the BCLK limitation, this sample reached its end quite early as we suspect that without a coldbug and higher BCLK capabilities this silicon would've lead to much higher clocks.



Intel Core i7 960 Results

Single Stage: -25°C

SuperPI 32M at 240x19 = 4560MHz (HT off)

SuperPI 32M at 235x21 = 4935MHz (HT off)

This Core i7 960 has been tested under phase-change rather quickly as I only wanted to know how the overclockability was in comparison to the Core i7 950 I tested before. In the end, it seemed that the chip is roughly 60MHz worse than the Core i7 950, but thanks to the 25x turbo multiplier it might be a bit better for Wprime liquid nitrogen damage coming next week. Do note that we were able to get 240MHz BCLK stable at only -25°C, which is quite a good result!

Step 1 - Learn the platform

If there's one thing I have learned in the past couple of years of extreme overclocking is that the foundation of a good session lies at the knowledge of your system. Too often I see people go straight from air cooling to liquid nitrogen and end up being frustrated that their system isn't overclocking all too well or not maximizing the system completely. As I don't have an LN2 sponsorship at this moment, I have to make sure that everything goes according to plan so that I'm not wasting any LN2 on worthless hardware.



The best advice I can give to a beginning extreme overclocker is to do one thing at a time: first make sure the base platform is working perfectly. This means pre-testing the combination of CPU, mainboard and memory and finding the sweet spots for voltage and performance. Also, know at what frequencies you can certainly boot/bench at so that in case something goes wrong, you can always go back to a safety setting. Another vital key in benchmarking, at least competitively, is that the first hour one should not be focused on getting the best score possible, but just getting a couple of runs through. This will give you a couple of base scores that you can compare to when benching for the real deal.

So, as my methodology prescribes I tested this Gulftown sample on phase-change cooling (~25°C) before going to liquid nitrogen. Thanks to the multiple visits to large overclocking events and spending hours in the past tweaking hardware for overclocking at subzero temperatures it's a bit easier to find the maximum frequencies, although a new architecture can always have uncovered problems. First of all, I tried to boot at the exact same voltage and frequency settings as I used for the Core i7 960. Not a common step, but ... I just wanted to see if it worked. And it did!

With a bit of tuning, I found this Gulftown to be capable of this 240MHz BCLK.



Sadly enough, that's where the fun stopped. I failed more than once to get this sample 3DMark Vantage stable, with all 6 Cores and 12 Threads enabled (6c 12t), neither voltage nor anything else helped me in that quest. Having heard so much good from the Gulftown overclockability in the past months, this was a huge slap in the face. I turned down the system and left it untouched for five days ... I was only going to test it again under liquid nitrogen, being almost certain it would be no good at all.

Liquid nitrogen, baby!

Having tested this particular sample on phase-change cooling which was, obviously, not such a satisfying session, we had mixed hopes when driving home the 25L dewar filled with liquid nitrogen.



On the one hand, there's the hope that this particular sample would turn out to be a killer chip, matching the other Gulftown overclocking results and going way beyond 6GHz 6c 12t stable through multithreaded benchmarks such as 3DMark Vantage and Wprime 1024M. On the other hand, there was the fear that this B1 would be pretty similar to what other people had reached with this revision of the Gulftown: no 6GHz, bad coldbug ranges and in overall disappointing.

First, let me show you the results of the test session and I'll comment on the results afterwards.

Test setup: Pictures





Test results

Coldbootbug: -85°C

Coldbug: -123°C

More on the temperature on the next page.

3DMark03 at 208x27 = 5617MHz (HT off): 113145

3DMark05 at 200x28 = 5600MHz (HT off): 43959

Aquamark3 at 215x27 = 5810MHz (HT off): 404584

SuperPI 1M at 200x28 = 5600MHz (HT on): 7.375 seconds

SuperPI 32M at 207x28 = 5803MHz (HT on): 6 minutes 36.469 seconds

Pifast at 215x28 = 6014MHz (HT on): 14.17 seconds

Wprime 32M at 207x28 = 5803MHz (HT on): 2.75 seconds

Wprime 1024M at 207x28 = 5601MHz (HT on): 1 minute 29.234 seconds

The most important information of this article you can find on the next page as we sum up our thoughts and findings regarding the overclockability of the Intel Core i7 980X. In the end, we cannot be unhappy with 6GHz benchmark stable since the Bloomfield only allowed us to go to 5.4GHz at most. A lot of these results are now entries in the top-20 overall rankings over at HWBOT, so ... it was a good first session.

At this point I'd also like to point your attention to the clock frequencies of the graphics card, the MSI R5870 Lightning. Although we are still in the test phase before the article, it's already quite impressive to see a solid 1425MHz core frequency. An article should pop online in the next couple of weeks.

Findings and some thoughts

First of all, it's not that easy to present a well-based conclusion (using valid arguments) about the overclockability of the Gulftown retail B1 revision after having spent only a 15L of LN2 in roughly 4 hours time. In fact, you must understand the consequences of what I am doing here: I'm trying to generalize the overclockability of one particular setup and form a conclusion that suits the entire range of Gulftown-based platforms, so I already ask you in advance to forgive me if the findings and conclusions turn out to be incomplete or inaccurate.

Moving on. As you might (or might not know), being into extreme overclocking requires me to be part of several extreme overclocking forums and websites to share information about overclocking. Many of my exploits are posted on forums a bit before they are turned into articles, because I prefer to hear the feedback from other users to check if my findings are somewhat accurate. When writing an article, however, a small forum post isn't enough, so I'll try to make it a bit more elaborative.

Keep the hero A0 or is it done with B1

The biggest question I have to tackle is one that have kept many extreme overclockers busy for the past couple of months: is the A0 revision really so much better than the retail B1? Comparing the results from the previous page to those you can find on HWBOT, you'd come to the conclusion that this is indeed true. However, we need to make a few sidenotes to this comparison before we jump to conclusions. First of all, as with the 32nm-based dual core codenamed "Clarkdale", the only variable that really seems to help the overclockability is the temperature. Much more than increasing the voltage, it seems that decreasing the temperature is beneficial for both getting the system more stable and reaching higher clock frequencies. What do I mean by this?

Increasing the Vcore helps a tiny bit to improve stability in the upper region of the "max clock zone", by which I mean that in case of almost 5.8GHz stable, adding a bit more Voltage will help to stabilize this 5.8GHz. However, adding a lot of voltage will not yield you 200MHz, certainly not at -110°C.

So, as already said, the operating temperature is the key for reaching high clock frequencies. And that's where the A0 versus B1 revision becomes a bit problematic. Because the A0 was an engineering sample only designed for system builders or mainboard manufacturers to check if their PCB design can handle a Gulftown, it does not have any security features built-in. For instance the lack of thermal or overcurrent protection comes in quite handy when overclocking and allows you to push the processor way beyond what it's designed for. Now, moan all you want, but adding these security features is a vital process of designing a new architecture, especially when they hit retail. Without these protections, normal end-users have an increased risk of damaging their €1000 processor even in normal test environments ... I don't think that's what we want. The upside of this story, for extreme overclockers at least, is that some A0 samples have no coldbug whatsoever, meaning they'll be able to operate at -196°C in theory, -185°C in practice. Given our "the temperature is vital"-finding, that's a huge advantage.

Overclocking B1 - Some tricky situations

Here's the tricky part however. With the B1 sample, I was able to reach 5.6GHz Wprime 1024M stable at only -110°C ... which is not that much worse than A0 at the same temperature and voltage level, according to several A0-overclockers. So, and this is where I might be wrong, I do believe that B1 revisions WILL match the early A0 overclocks, but only if you find one that is capable of running the very low temperatures. Not many have been tested so far, so it's quite difficult to come to a conclusion, but we certainly hope this will be happening in the future. In any case, I most certainly invite Intel to meet up with overclockers and check if there's any correlation between stepping and lowest bench-able temperature. Preferably myself *wink*.

In addition, I have to add that A0 revisions tend to die quite quickly and have an inferior memory controller on board.

How did the UD7 do?

So, was this session all happy-happy pink roses? No, because I experienced several other problems, not all related to the processor. For one, it seemed that the Gigabyte X58A-UD7 was not capable of handling either the extreme current or the extreme temperatures. Let me explain.

While I was indeed reaching 5.6GHz 12-thread stable, sometimes for no apparent reason the motherboard shut down and locked itself in the "88" position, meaning "unable to start POST". Powering off the PSU, let the board rest for a couple of seconds and powering everything back on certainly solved the problem (even bios settings were saved), but the sudden shutdown combined with the inability to commence post is a clear indication of current problems.

As for understanding the extreme temperatures issue, it's important to know the difference between coldbug and coldbootbug. The coldbug is the absolute temperature at which the component fails to operate. The coldbootbug is the absolute temperature at which the component fails to boot up. The coldbootbug is generally a bit higher than the coldbug, because it seems that components have no problem continue to work at a given temperature, but have more problems going from idle to load at a given temperature. To give an example: the coldbootbug may be around -100°C, but when in windows the coldbug may be -125°C. The issue in this case is that the coldbug and coldbootbug behaved in a very atypical manner. It's a bit hard to explain, but in general it works like this:

- Coldbug => board shuts down completely and doesn't commence POST (88)
- Near coldbootbug => board starts POST, but fails regularly (C1-C2-C3-shutdown-C1-...)

In this case however, the board shut down at exactly -123°C, but it didn't behave like a coldbug should as it returned to C1 directly, not 88. Dropping the temperature even more, I found the coldbug-like behavior to appear at -155°C, which makes me to believe that the processor is actually capable of much lower temperatures and, thus, higher frequencies. This is something we need to figure out in the next couple of weeks when testing other mainboards!

Last couple of things ...

Another subject I have to tackle is an issue with VTT voltage scaling, but this depends from sample to sample. I was not able to set a VTT voltage higher than 1.5V or the mainboard would refuse to boot. Although the memory controller quality of the B1 revision is a lot better than the one of A0, with 1.495V it's unlikely we'll see clock frequencies in the 5GHz range. Other people report the same issue, but at different VTT voltage levels.

One last thing for the extreme overclockers under us: this architecture consumes liquid nitrogen like crazy, especially when stressing all 12 threads in multi-threaded applications like Wprime 1024M. At first sight maybe a bit strange given that the TDP of a 32nm 6-core (980X) is equal to that of a 45nm 4-core (975), namely 130W, but you have to know that in contrary to the Bloomfield platform, this architecture is actually scaling with voltage. Whereas our beloved quad cores maxed out around 1.6 to 1.65V, the Gulftown can easily go up to 1.8 to 1.9V, which is exactly where the difference is located. The issue of the temperature was noticeable even while using the solid copper Kingpincooling Dragon F1EE, known for its mass. To give you an example: when I start the Wprime benchmark at -110°C and a container 1/2 filled, the temperature was raising 0.3°C per second. Even with a fully filled container, the temperature drop was still noticeable as ~ 0.15°C increase per second.

Evaluation

An evaluation before we have even seen the performance results of this product? Before we have seen the comparison with the previous generation of hardware? Are ... you ... mad?!

As per definition, yes, we are mad. Also, we did run our performance comparison benchmarks as well as check out the other 43 reviews of the Gulftown and they all lead to the same conclusion: any performance comparison is pointless. In fact, writing the evaluation hasn't been this easy since a long time: if you have no problem spending €1000 to get the best performance, then you need no comparison charts to know that the Gulftown will be delivering the highest performance. If you take price into account, then no, you will be better of with a different product (be it Bloomfield or Phenom).

For all those readers who are waiting to find a "discuss this article"-hyperlink and rant on how expensive this product is and how Intel should make this cheaper to sell more: no, they shouldn't. Yes, it's justified to ask this much money for a product. No, it's not unfair. Yes, in the future we'll see the same technology but at a lower price. As I have said on some enthusiast forums: the point of this product is not to sell en mass, but to showcase the upcoming technology. The point is not to make a huge sale, but to make people who want to have the best and are willing to pay for it, be able to get it. Any price-argument is non-valid, in my humble opinion. We know it's expensive, so deal with it.

Intel Core i7 980X recommended For


As we have been testing the Gigabyte X58A-UD7 quite intensively over the last couple of weeks, we can say that the board definitely allows the end-user to get the most out of his processor, especially on LN2 cooling. Although there were a couple of problems with the Gulftown, we have to admit that with our two Bloomfield samples the board served us perfectly and has lead us to a couple of nice overclocking scores (Top-20 HWBOT). Allowing upto 240MHz BCLK with phase-change is definitly enough to max out most of the Core i7 Bloomfield series. Having had my share of X58 mainboards on the test bench (~ 14 different models), I do have to admit that this board has given me the most pleasure already, so top notch. Do note that it comes at quite a price as well.

Gigabyte X58A-UD7 recommended For


Conclusive thoughts

Has the Gulftown impressed me? Yes. Will I be happy to continue to use this product in my future benchmark sessions? Firm yes. Will I buy one to put in my 24/7 desktop system? No way. I think this pretty much sums it up: the Intel Gulftown architecture is so powerful, there's so much performance coming from it that I just can't deny that at the moment it's the only product I want to use in extreme overclocking sessions. Realistically speaking, however, I have to admit that because the price not a lot of people will be eager to buy this product, but for those people there will be a Core i7 970 coming soon which should retail at a much lower price (est. €450/500). When that product comes to our lab, we will have a more extensive 'normal' benchmark procedure as then price-arguments will become valid again. And, with Thuban (AMD's 6-core) coming next month, we might actually see an interesting price-war between the two or three 6-core products. Interesting for sure!

We will try to continue our extreme overclocking testing in the next couple of weeks and try to find out if there's a better mainboard or CPU stepping out there for the overclockers under us. After, if you are intended to buy this product for extreme overclocking, you want the first sample you buy to be of the best available stepping!

To end with, I'd like to thank the following people for providing the hardware and support. Without these people Madshrimps (and I) would never be able to provide you this content.

INTEL Belgium (Kristof and Jan): for providing the Intel Core i7 980X, as well as the previous Intel samples
Gigabyte Benelux (Bernice): for providing the Gigabyte X58A-UD7 sample
OCZ Technology (Tobias): for providing the batch of OCZ Freeze cooling paste
TONES (Manu): for always ready to provide hardware where necessary



For more overclocking results of the Core i7 980X and many other hardware, please visit HWBOT.org, the worlds largest OC database.



On the next two pages you can find performance results of the Core i7 980X compared to the Core i7 960, both clocked at 4.8GHz to see the difference in performance between both architectures. Since this is a Gulftown article, we have not included an AMD Phenom in the testing procedure for the simple reason that we don't want to compare apples to pears. When the Thuban arrives, it will make for a nice shoot-out.

The comparison charts - part 2

Last set of the comparison charts. The conclusion in fairly easy, I guess: from the moment the application scales with multiple cores, the 980X will take over the head of the charts. In single threaded applications in which cache size plays a role (eg: Winrar), the 980X will be a tad better due to the increased cache size. In single threaded applications where only raw CPU computing force is needed, the Bloomfield architecture can match the Gulftown's performance.