facebook rss twitter

Review: ASUS P7H57D-V EVO motherboard for Intel Core i3 and i5 chips: H57 fully loaded

by James Smith on 27 January 2010, 09:35 3.4

Tags: ASUS P7H57D-V EVO (H57), ASUSTeK (TPE:2357)

Quick Link: HEXUS.net/qavol

Add to My Vault: x

System setup and notes

Motherboard ASUS P7H57D-V EVO Intel DH55TC
Price £149.87
£81.72
Chipset Intel H57 Intel H55
CPU Intel Core i5-661 LGA1156
(3.3/3.6GHz, 0.5MB L2 Cache + 4MB L3 cache, dual-core + HT)
BIOS revision 0401 (12/02/2009)
0027 (11/19/2009)
Mainboard software Intel Inf 9.1.1.1020 + Matrix Storage Manager 9.5.0.1037
Memory 4GB (2 x 2GB) Corsair Dominator GT CMD8GX3M4A1600C8
Memory timings and speed 9-9-9-24 2T @ DDR3 1,338 9-9-9-24 1T @ DDR3 1,333
Integrated graphics Integrated GMA HD 256MB (900 / 1338) Integrated GMA HD 256MB (900 / 1333)
Integrated Graphics driver 15.16.4.64.2008_PV (8.15.10.2008) 
Discrete graphics HIS Radeon HD 5770 1GB (850 / 4800)
Discrete graphics driver Catalyst 9.12 (8.681-091124a-092499C-ATI)
Disk drive Seagate 1,000GB SATAII (ST31000528AS)
Operating system Windows 7 Ultimate (64-bit)
PSU Corsair CX400W

Tests

Benchmarks

SiSoft Sandra 2010 (16.11), float buffered
ScienceMark 2.0 memory latency
HEXUS.PiFast to 10m places
StaxRip X.264 + AAC encoder - two-pass (both passes recorded)
CINEBENCH R10 multi-CPU render 64-bit
POV-Ray 3.7.0 beta 34 64-bit
wPrime 1024m

HDTach - SATA average read speed
HDTach - SATA burst speed
HDTach - USB average read speed
HDTach - FireWire average read speed

Far Cry 2 v1.03 - 1,024x768 very-high-detail 0xAA
Enemy Territory: Quake Wars v1.5 - 1,024x768 low-detail
Enemy Territory: Quake Wars v1.5 - 1,680x1050 high-detail

Power-consumption tests
Overclocking tests

Testing notes

The first two Clarkdale-compatible motherboards we've received are at vastly different price points and in different form factors, thus the Intel DH55TC has been included purely to provide reference performance numbers.

When testing a motherboard one of the most important aspects is to evaluate its stability, so what better CPU to use than one with a relatively-high TDP. The Core i5 661 fits this bill nicely, which despite its slightly-lower clock speed of 3.3GHz compared to the i5 670's 3.47GHz, has a higher TDP of 87W versus 73W.

Additionally, the clock speed of the built-in Intel HD graphics is higher than that of the Core i5 670. On this note, a Radeon HD 5770 makes a good choice if you decide to upgrade to a discrete graphics card solution - we've found it offers decent playable DX9 and DX10 framerates in most of our benchmarks at 1,680x1,050, and provides DX11 compatibility to boot.

The final component choice which is likely to have a noticeable effect on performance is the memory. It might seem a bit odd at first, to choose a relatively expensive DDR3 1,600MHz kit to test with on a mid-range motherboard, but this choice actually makes a lot of sense. For the majority of benchmarks we run these at DDR3 1,333MHz with the same CL9 timings you're likely to find on budget 1,333MHz kits, therefore performance should be the same.

However when it comes time to testing overclocked performance, we can happily keep the same modules installed and run them at their rated CL8 DDR3 1,600MHz speed. A couple more benefits of using this kit are to check the X.M.P. functionality works in the BIOS and also to prevent any memory-related limitations when finding the overclocking limits of the board.

All of the 2D benchmarks are run using just the integrated graphics of the Intel Core i5 661 CPU.

However, apart from the relevant 3D benchmarks and power numbers, the Radeon HD 5770 graphics card is installed when running the overclocking tests in order to prevent any GPU-related limits from holding things back.

Issues

During our testing we found that when saving and later restoring BIOS profiles, neither manual CPU ratios or manual DRAM timings were restored correctly.

The CPU ratio would restore as what was preset in the X.M.P. profile (x21) , whereas the CAS-Latency, RAS-to-CAS delay, RAS precharge time, RAS cycle time, and DRAM timing mode (command rate) would always restore as 'auto' timings instead of the 9-9-9-24 2T or 8-8-8-24 2T manual timings specified.

On another note, it wasn't possible to run exactly the same memory command rate on the Intel DH55TC motherboard, or run the overclocked numbers. No matter what we tried, we couldn't get access to either manual memory timings, or CPU ratio adjustments.

After a brief investigation, it was discovered this is most likely due to a known limitation with the current board revision regarding memory voltage adjustment. Thus we think this whole section of the BIOS has been deliberately locked-out.