When I was made redundant by Marconi I decided to use some of the money to buy the big brother of the AMD Duron - the AMD Athlon "Thunderbird". The Athlon has improved performance over the Duron at the same clock rate primarily due to the faster 256k cache (compared to the Duron's 64k).
When Intel introduced the "coppermine" Pentium III they had the fastest and best performing desktop processor around. However, AMD soon caught and surpassed them with the Socket A Athlon "Thunderbird". Since then it's been a ding dong battle - Intel introducing the Pentium 4 and AMD producing faster versions of the Athlon that were beaten only by the P4's DDR (remember - data clocked on both edges) and RIMM memory support. Now there are AMD supported chipsets that handle DDR memory (such as Via's KT266a) Intel again have some catching up to do.
For instance, my Athlon 1000@1466 outperforms a P4 1.6 GHz as demonstrated below. The only area where the P4 excels compared to my current system is in the memory tests with DDR against standard SDRAM.
I chose the fabled "AXIA" version of the Athlon. This marking is the 3rd line down on top of the processor core and the "AXIA" marked versions proved to be very over-clockable with nearly all making 1.4 GHz or better. The "AXIA" is a 100 MHz FSB version.
WCPUID & SiSoft Sandra results :-
Achieved with an FSB of 133 and a multiplier of 11.
Designed for AMD Athlon beyond 1.4GHz
High speed 2 ball-bearing fan (powerful, compact, and quiet)
Cold forged Aluminum and Copper for enhanced heat transfer
Proprietary clip design allows easy attachment with the proper amount of force for the right contact
Compatible with Socket 462 (Athlon and Duron)
Speed 4800 RPM
Max. Air Flow 21.19 CFM (Min. 19.07)
Very quiet compared to the GlobalWin FOP38
With this heatsink I managed to get the Athlon stable at 52°C under full load and 1.98 voltage. I may change the fan to the Delta 38 6800 RPM version used on the FOP38 to get the temperature down a little bit.
Leadtek have a reputation of producing some of the best nVidia chipset based graphics cards and regularily produce cards with "overclockers" in mind. You can tell this by the presence of the hugh heatsink (which covers both sides of the board because the memory is on both sides) in the photo above. The heatsink also integrates 2 fans as opposed to the normal one to cool both the chipset and memory (note that the A250 LE only has the one fan and is based upon the Ti4200 - see the table below). The GeForce4 chipset used on this card is the latest in the GeForce series following on from the successful GeForce2 and GeForce3.
The GeForce4 chipset is accompanied by varying peripheral components to produce 6 main flavours (all based upon nVidia reference designs) :-
| Default Chip and Memory Speeds of GeForce4 Based Cards | ||||||
| Card | MX420 | MX440 | MX460 | Ti4200 | Ti4400 | Ti4600 |
| Default Video Clock Speed (MHz) | 250 | 270 | 300 | 250 | 275 | 300 |
| Default Memory Speed (MHz) | 166 | 200 | 225 | 225 | 275 | 325 |
The main difference between the MX and Ti series (apart from the reduced memorty speed) is that the Ti series include nVidia's nfiniteFX II Engine. To quote nVidia - "The NVIDIA nfiniteFX II Engine incorporates dual programmable Vertex Shaders, faster Pixel Shaders and 3D textures. The nfiniteFX II Engine gives developers the freedom to program a virtually infinite number of custom special effects to create true-to-life characters and environments." If you want more detail visit the nVidia website.
All of these chipsets can use DDR (Double-Data Rate) memory where data is clocked on both rising and falling edges of the clock as opposed to one edge only on SDRAM. This means that the memory speed is usually quoted as double the figure shown in the above table - ie, 275MHz default memory speed is usually quoted as 550MHz.
nVidia release drivers on a regular basis and there is a feature built-in called "Coolbits" which allows you to overclock the clock and memory speeds. If you do run the video and memory clock at higher than default speeds be aware that this means more heat will be generated and therefore more cooling may be necessary.
With the Leadtek A250 series this is made easier to achieve by supplying their "WinFox" utility with the card which allows you to overclock it (note - these are my default settings) :-
Below are the results obtained using the popular 3DMark2001 SE test program and nVidia's 40.72 reference drivers :-
I'm quite pleased with these results considering I'm still using the old Athlon rather than the newer Athlon XP and my motherboard only supports PC133 DRAM (single data rate) memory as opposed the newer PC2100, PC2700 and PC3100 DDR DRAM. At 315/650 the card settings are just above the default for the Ti4600 and the only limitation is the memory speed. At these settings there were no faults during the graphics testing - known as "artifacts".
Stock 10Base-T affair.
Used with NTL's broadband 512kbit/s service.
DOCSIS 1.0 certification by CableLabs and advanced MIB support ensure compatibility with DOCSIS-based service networks
Field-upgradeable software minimizes truck rolls and reduces product life cycle costs
High down-stream data rate up to 38 Mbps, as well as ability to forward 14,000 64-byte packets-per-second, provides full line-speed data
Supports up to 16 PCs for workgroup and small office networks
IGMP Proxy support enables efficient control and delivery of streaming media multicast services
Compact size and simple installation with a small, unobtrusive external power adapter
Full product life cycle support means worry-free operation
Finally got around to replacing the old HP8250i 24x 4x 4x. The key features for this one were speed, price and the fact that it was BURN-Proof (which virtually eliminates bad CDR's).
Bought to replace the Seimens Nixdorf MCM2103 21" which finally became annoying with screen flicker and power problems. I decided upon quality rather than cost this time with a flat screen capable of supporting high refresh rates. My current default set-up is 1152x864 @ 100Hz
FLATRON 915FT+ |
Auto-synchronous support from 640 x 480 up to 1880 x 1440 @ 70 Hz NI
Flicker-free up to 1600 x 1200 @ 85 Hz
Digital (microprocessor-based) On-Screen adjustments, soft sensor controls
Internal screen coatings further reduce reflection
The flat screen surface eliminates fatiguing glare
Innovative design offers more accurate color, better sharpness
No annoying horizontal tension wires, as seen on other flat monitors
Built-in 1 up 4 down USB Hub
PrimaScan Colorado 2600u Scanner |
Optical resolution: 600 x 1200 dpi
Hardware Supersampling Resolution: 1200 dpi
Enhanced Resolution (interpolated): 2400 dpi
Scanner Bit Depth (colour): 42-bit internal (over 4.3 trillion colours)
Scanner Bit Depth (grayscale): 14-bit internal (16,384 shades of gray)
24-bit colour, 8-bit grayscale
Scanning area: Up to 21.6 cm x 29.7 cm (A4/letter) (8.5”x11.69” )
Scan Method: Fast single pass
Interface: USB Port
Driver: TWAIN with advanced scanning interface
Light Source: Cold Cathode Fluorescent lamp
Dimensions: 45 x 29 x 8.6 cm (17.75” x 11.4” x 3.4”)
Weight · 2.5 kg (5.6 lbs)
15.3 GB ATA-100 - Maxtor DiamondMax Plus 45 Ultra DMA 100 - 51536H2, 8.7ms seek time, 7200 RPM
Formatted as:
2.99 GB (C:) - WinME & Applications
1.00 GB (D:) - Swap file
3.49 GB (E:) - Downloaded applications, updates, patches, drivers, MP3's
2.82 GB (F:) - Microsoft Links 2001
3.96 GB (G:) - Emulators (Amstrad CPC, Sega Megadrive, M.A.M.E., Nintendo 64)
20.0 GB ATA-100 - Maxtor DiamondMax Plus 60 Ultra DMA 100 - 5T020H2, <8.7ms seek time, 7200 RPM
Formatted as 9.49 GB (H:) and 9.48 GB (I:) - both for Games
8.4 GB ATA-66 - Maxtor DiamondMax VL17 - 90871U2, 9.5ms seek time, 5400 RPM
Back-up drive for the C:, D: and E: partitions on the 15.3GB drive. Back-up's are made using the excellent Norton Ghost 2001 (see a review here). This drive isn't normally installed in the machine (because of the obvious drive letter conflict) so I use a removable IDE caddy so I can insert it when needed.
Removable IDE Caddy |