EVGA P55 FTW - It's Here...Pictorial Preview
by Rajinder Gill on August 10, 2009 11:07 PM EST- Posted in
- Raja's Ramblings
Sorry folks, we actually have another shameless P55 motherboard picture preview for you. I know Gary thought his ASRock article was the last of them, but EVGA released us this evening to discuss their flagship P55 (E657) FTW (For The Win) product. We know you are urging us to break rank at this point and post up some heavy overclocking benchmarks and performance comparisons against the i7 platform, but there is little chance of getting away with anything other than a few pictures to look over until the platform embargo lifts.
We are not quite sure at this point of how many users will be attracted to higher cost i5 boards given the fact that that total platform cost overlaps the cost of entry-mid level i7 offerings. However, this has not deterred any of the big players from putting together a top-end i5 package based upon a combination of features they think will reel in customers. Early pricing lists of i5 boards are putting the top-end offerings between the $200~$250 mark, this one is no different with an estimated MSRP of $250 (there will be a few sub $200 models offered as well as an NF200 'Classified' variant of the board we're previewing today).
Any board following in the footsteps of the EVGA X58 Classified has a lot to live up to. Early overclocking tests on the i5 platform have revealed a similar trend to what we saw on the X58/i7 motherboards when used with sub-zero cooling, with the top-end EVGA P55 boards ahead of the competition by the slimmest of margins as we move into the final tweaking stage before full retail release.
It is not a massive gap at this point, we are talking a 3-5 BCLK difference between this board and the ‘best of the rest’ given the same hardware configuration and voltages. For the more conventional user (those using air and water cooling), our take is that you won’t find any overclocking advantage comparing one enthusiast board over another, assuming functionality is completely up to speed.
There are a number of unique hardware features that EVGA hopes will help keep their top line boards ahead of the pack when really pushed by an enthusiast. EVGA has gone the extra mile by providing onboard hardware components that make fine adjustment compensation for all major signal lines possible among other items. Team these functions with sub-zero boot-up & current limit manipulation jumpers and we begin to see that EVGA is making a full assault for the P55 overclocking crown.
Companies such as EVGA, DFI and MSI continue with a purist approach to PWM design for all major power rails. We happen to sit with the camp that believes in the importance of fast transient response and low noise attributes first. The proof is in the pudding and we’re eager to compare several boards side by side to see if any additional overclocking headroom can be realized that supports the ‘more phases=unconditionally better’ theory.
Initial in lab tests indicates that the only thing that really matters from a user perspective is that the given PWM design is capable of pumping out the required current comfortably and cleanly. The rest comes down to heat management and ease of cooling; anything more seems to be a matter of marketing rather than real substance.
Other than costs, we cannot see a reason why it is not possible to keep a reasonable phase count by using suitably sized MOSFET’s with a decent heat sink and capable buck controller. We think it would be of more use to all concerned if companies would simply inform all of us how much current their PWM circuits can supply rather than tell us about XXX number of phases, which on its own does not tell us anything at all.
In the case of the E657 FTW, CPU power is supplied by a 600W peak current capable 12-phase supply (more than i5 will ever need). VTT and memory power is supplied by 2 and 3 phase supplies respectively, both using the same model of FET’s as used for the CPU. You’ll also notice that EVGA has added socket 775 mounting holes at a slight angle from the 1156 holes, a welcome addition for those of you with a good 775 heat sink or those of you concerned about needing new socket mounts for those trusted LN2 pots and phase change coolers.
The rear panel contains 6 USB ports, PS2 keyboard port, 2 ESATA ports, 1x1394a port, dual Marvell Gigabit LAN ports and the audio panel driven by the popular Realtek ALC889 HD codec. A CMOS clear button is provided on the rear panel for ease of use when the board is mounted in a case.
A Hex postcode display is available for debugging purposes together with an output connecter for EVGA’s OCP (overclocking panel) which helps users work around the somewhat quirky boot-up voltage requirements of sub-zero cooled processors. An additional 6 USB ports are supported via internal headers, plus an extra 1394a header.
Three onboard BIOS chips are present allowing users to store different BIOS revisions for testing purposes or as a failsafe backup. BIOS chip selection is made possible by a simple three-position slider switch placed at the lower left corner of the board.
Looking over the slot layout; we have one PCIe 2.0 x1 slot, two PCI slots, and three PCIe 2.0 x16 slots providing x16/x4 or x8/x8/x4 lane allocation (depending upon usage). Seven (1 CPU + 6 additional) fully controllable fan headers are placed in various locations, five of which are easily accessible at all times regardless of component loading.
Onboard power, reset and CMOS clear buttons are placed at the lower edge of the board for those of us that like to test and use boards on an open testbed (the reset button doubles up as a flashing LED to show SATA activity).
That about rounds up today's quick preview, the official release is not far off so please bear with us...
54 Comments
View All Comments
strikeback03 - Tuesday, August 11, 2009 - link
I wonder why Intel felt the need to create another set of mounting holes. The 1156 holes are what, less than a mm further out than the 775?GeorgeH - Tuesday, August 11, 2009 - link
What's the reasoning for the molex connector by the x1 slot?ianken - Tuesday, August 11, 2009 - link
I;d like to know that too.Alos, is the back of the CPU soicket clear of clutter? Every evga board I've used had crap back there that prevented the use of after market coolers with back plates.
AznBoi36 - Wednesday, August 12, 2009 - link
I remember these. I had them on my 680i LT SLI board also. Seems to be a "EVGA" thing.Rajinder Gill - Tuesday, August 11, 2009 - link
The molex connecter is to augment PCE/e slot power when using a couple of heavily overclocked GPU's. It's usage is not a necessity unless you're really pushing the GPU's hard.Cooler backplates designed for use with i5 boards should be fine. The only infringement into the mounting area is the presence of thru hole cap leads which are a requirement gievn the need for bulk capacitance close to the CPU socket (all boards have them - some a lot more than others). Vendors designing coolers for socket 1156 should already be aware of this and be using suitable brackets for upcoming coolers.
later
Sagath - Tuesday, August 11, 2009 - link
Is this a chipset limitation or a choice by EVGA to try and push those with SLI to X58?Also, its been said a million times before, and needs to be said again; Let PCI die.
Nice lookin' board. I guess they listened to the feedback from the X58 SE. I loved the look, just didnt have the features I wanted.
MadMan007 - Tuesday, August 11, 2009 - link
Chipset limitation: 16 lanes from the CPU northbridge and 4 from the P55 chipset.I'm not too bothered by PCI on full ATX boards if they are layed out right to take in to account those who actually use that many cards. I do however wish that PCI would die faster on mATX boards.
Casper42 - Tuesday, August 11, 2009 - link
Just to clarify, P55 supports up to 8 PCIe lanes.But some of those will be eaten by Gigabit LAN, additional SATA Controllers, etc.
The beginning of the article mentions that the "Classified" version of this board will use an NF200 chip, so you will be able to do:
16-16-4 for full bandwidth to both cards or possibly even 16-8-8 for Triple SLI.
tomoyo - Tuesday, August 11, 2009 - link
You know, PCI really can't die unless peripheral makers actually make PCI-E CARDS. It's complete failure how there's a total lack of pci express 1x cards for a ton of things. If there was a good marketplace of every type of device in PCI-E style, we'd be all set to kill PCI.MadMan007 - Tuesday, August 11, 2009 - link
Perhaps you haven't looked at add-in cards in quite some time but just about every type of add-in card is available in PCIe nowadays. The main reason for PCI seems to be for those carrying over cards they already own, much like IDE which people like to hate upon more (some older optical drives that are IDE are better for certain things than new ones.) I don't understand the IDE hate though, it takes up a minial amount of space and doesn't really take away from what you can do with the PC unlike expansion slots. It does look like this board has pads for an IDE port though ;)