- The Motherboard Continued
While not a huge requirement on the desktop platform, SCSI controllers are pretty much second nature to the server environment, in fact, SCSI in servers outnumbers IDE 3:1. SCSI has proven to have a much higher sustained transfer rate, as well as being able to have more than two devices per channel does give SCSI the advantage. The Adaptec AIC-7902W Ultra320 SCSI controller is built directly onto the motherboard with the chip interfacing with the PCI-X bus at 100MHz, giving it a high speed connection to allow data to pass to the rest of the system without the constraints of the 33MHz PCI system. Two 68pin Ultra Wide SCSI ports are located at the bottom right of the motherboard. Each port can support seven drives, that's a total of 14 drives on the SCSI controller, with RAID 0, 1, 0+1 and RAID5 support; you will certainly have redundancy with this setup.
Serial ATA support is a must have for any new motherboard - server or desktop. While the ICH does support two native Serial ATA ports, Gigabyte simply felt that wasn't enough. The tried and true Silicon Image Sil3114 four channel Serial ATA controller is built on to the PCI bus to give four extra Serial ATA ports.
Since the ICH only supports four native USB 2.0 ports, the VIA VT6202 four port USB 2.0 controller chip makes a return, this chip was big on Intel motherboards that didn't support USB 2.0 during its initial introduction. Today it returns to add four extra USB 2.0 ports, a total of 8 onboard.
Gigabyte has in the past shown us that CPU's are ever changing and requiring more and more power to function in overclocked situations. Even before Prescott hit, Gigabyte was introducing dual power modules to boost their boards from three phase power to six phase power. Traditionally it has been through a riser card added into a VRM slot. This time Gigabyte has come out and simply integrated the extra three phases directly to the motherboard. Under the heatsink and fan are six phase voltage regulators, which come especially handy if you have a power hungry Prescott, great for stability in server environments when stable systems are a must have.
Though you may find this odd to see, the Gigabyte 8KNXP Ultra64 features all of the overclocking features its desktop motherboards do which is very unusual for a server classed motherboard, showing Gigabyte tend to push this motherboard at the more performance users who want the lot.
Under the Frequency/voltage control menu you find all the overclocking options including FSB speed, DRAM ratio, AGP/PCI/SATA lock, Vagp, Vdimm and Vcore.
FSB settings are from 133MHz all the way up to 355MHz in 1MHz increments, while 355 is almost impossible to reach, it is good to see a full selection. DRAM ratios are selectable depending on what speed your CPU is, but none the less you will get speeds for 266, 333 or 400MHz.
Locking the AGP/PCI/SATA to 66/33/100MHz respectively keeps the main buses in check, so your add-in devices aren't the cause for a poor overclock.
Voltages are a mixed bag. CPU voltage is selectable from 0.9v up to 1.75v in 0.025v increments. Doesn't matter if it's Prescott or Northwood, they all have the same settings, though pushing the Prescott above 1.55v is a bad idea. DRAM and AGP voltages in the past have been very limiting and again Gigabyte isn't listening to the demands of the users. You can only set your DRAM and AGP voltages to a max of +0.2v above standard, this means DRAM max is 2.8v and AGP is 1.7v. DRAM - voltages to be competitive need 3.2v these days, as modules are designed to handle up to 3v now.
That said our maximum FSB we achieved was 253 MHz. When trying higher, the DRAM would simply cause system resets - more DRAM voltage is needed.
PRICING: You can find products similar to this one for sale below.
United States: Find other tech and computer products like this over at Amazon's website.
United Kingdom: Find other tech and computer products like this over at Amazon UK's website.
Canada: Find other tech and computer products like this over at Amazon Canada's website.
- Gigabyte 8KNXP Ultra 64 - Page 1 [Introduction]
- Gigabyte 8KNXP Ultra 64 - Page 2 [Specifications]
- Gigabyte 8KNXP Ultra 64 - Page 3 [Features]
- Gigabyte 8KNXP Ultra 64 - Page 4 [Features Continued and Overclocking]
- Gigabyte 8KNXP Ultra 64 - Page 5 [Benchmarks - Test System Setup and Sandra 2004]
- Gigabyte 8KNXP Ultra 64 - Page 6 [Benchmarks - PCMark04]
- Gigabyte 8KNXP Ultra 64 - Page 7 [Benchmarks - 3DMark03]
- Gigabyte 8KNXP Ultra 64 - Page 8 [Benchmarks - Aquamark3]
- Gigabyte 8KNXP Ultra 64 - Page 9 [Benchmarks - Halo PC]
- Gigabyte 8KNXP Ultra 64 - Page 10 [Benchmarks - Jedi Knight II]
- Gigabyte 8KNXP Ultra 64 - Page 11 [Benchmarks - Unreal Tournament 2003]
- Gigabyte 8KNXP Ultra 64 - Page 12 [Conclusion]
- We at TweakTown openly invite the companies who provide us with review samples / who are mentioned or discussed to express their opinion of our content. If any company representative wishes to respond, we will publish the response here.
Latest News Posts
- Samsung has 3.5-inch 120Hz OLED: next-gen VR panels
- Samsung teases prototype 1.96-inch 4K LCD for VR HMDs
- Exploding Beats headphones aren't Apple's fault
- Samsung: Galaxy S8 is 2-for-1 in the US only right now
- NVIDIA's purported Volta-based TITAN Xv spotted
- ASUS RT-AC1900p Wireless Router Review
- ASUS RT-AC1900p Wireless Router Review
- GA-Z97X-UD5H-BK with SSD M.2 Plextor M8PeG 256gb. low speed after shut down
- Prey Review: Dark Stars, Darker Thoughts
- Alien Covenant Movie Review
- Western Digital sets new standard with latest generation in popular HGST-brand Ultrastar SAS SSD family
- AMD raises expectations for server performance, unveils EPYC processor brand for the datacenter
- HighPoint's SSD7101 PCIe board-sized drive series integrate Samsung 960 NVMe SSDs to deliver groundbreaking performance over 12GB/s
- Micron accelerates all-flash storage speed, performance and value with new flexible petabyte-scale enterprise data center solution
- Qualcomm Snapdragon 660 and 630 mobile platforms drive advanced photography, enhanced gaming, integrated connectivity and machine learning