Firingsquad and others.... here is a clip


[ Follow Ups ] [ Post Followup ] [ CPU-Central Message Board ] [ Home ]

Subject: Firingsquad and others.... here is a clip
Name: me
Date: 7/24/2001 4:08:01 PM (GMT-7)
IP Address: 63.208.185.54
In Reply to: WHO says it is the norm posted by ludicrous
Message:

every website that does a comparison...

For instance. Firingsquad....

In integer performance (CPUMarks), the KT7A-RAID (and the VIA KT133A chipset) performs within 2% of the Gigabyte GA-7DXC AMD-760 DDR motherboard. This is pretty impressive considering the lower cost of the KT133A upgrade, but keep in mind that the gap will increase as the processor clock speed scales higher. Even the KT133-based KT7-RAID scores well here, despite its slower system bus.

Thats with a 1.2Ghz

In disk tests based on real-world applications, the 760 and KT133A motherboards overtake KT133. The KT133A/HighPoint combination offered with the KT7A-RAID is just able to outpace the 760 Gigabyte board by 3% in Content Creation Winstone 2000. In Business Winstone 99, the KT7A-RAID and Gigabyte GA-7DXC perform within 1% of each other, a dead heat.

Our synthetic CPU tests hardly discern the difference between the KT133A and 760 motherboards represented here. But for that matter, the KT133 chipset is right up there with them as well.

Likewise in 3D instruction tests, all three chipsets represented here perform extremely close to one another. These tests just can't discern the difference between the different bandwidths available from each solution.

We finally see the difference double data rate memory can make in SiSoft's memory bandwidth tests, we witnessed an 8% difference between the GA-7DXC and the KT7A-RAID in integer memory performance, and a whopping 22% difference in floating point memory bandwidth. We also see the KT7-RAID hanging with its younger brother in floating point tests, but trailing by 8% in the ALU/Memory Bandwidth test.

In sisoft hard drive
While the GA-7DXC hangs in there with the KT7-RAID, the KT7A-RAID trails by a surprising 6%. We were a bit surprised by this result, but the results repeated themselves.

First of all, keep in mind that with OfficeBench testing, a lower score is better. OfficeBench runs through a set of scripted sequences in Word, Excel, and PowerPoint, measuring the time it takes to perform actions within each program. At the end it runs them all multitasked and records the total time. Depending on the test, the KT7A-RAID and GA-7DXC perform the fastest, although as you can see, the difference is within one second (Although statistically, that second can add up to five percentage points or more).

The 760 and KT133A motherboards perform within three percentage points of each other in 3DMark 16-bit testing before evening out at 1024x768x16. The edge goes to the GA-7DXC 760 motherboard. Despite this, MadOnion's CPU benchmark gives the nod to the KT133A by two percentage points. It seems as if this result should have been the other way around.
The KT7A-RAID closes the gap in 32-bit testing, where the video card becomes more of a limiting factor.

In UT, we'll begin to really see the benefits of the 266MHz system bus offered in the KT7A-RAID and GA-7DXC, as well as the DDR memory itself.
As you can see, the faster system bus boosts performance in UT by about 7.5%. However, the DDR memory itself only adds an additional percentage point of performance. At 800x600, you still see the 7% difference between the 266MHz boards versus the 200MHz KT7-RAID, while the DDR memory still adds about a percentage point.

As the resolutions scale higher, the video card plays more of factor and all the systems begin to even out.

MDK2 doesn't change the situation much from UT. The KT7-RAID is able to slightly close in (now 6%), while the KT133A continues to represent itself well against 760.

In Quake 3 we see the DDR 760 motherboard open up the gap a little from KT133A, now at 3%. Likewise, the KT7-RAID trails its newer counterparts by a slighter greater margin. Again, keep in mind that as the CPU clock speed increases, we'll see the gap between all three chipsets continue to increase.

Game, set, match.....



[ View FollowUps | Post Followup | Main ]



Follow Ups:



Maximum of 100 messages displayed.

Post a Followup

Name:
E-Mail:

Subject:

Comments:

Optional Link URL:
Link Title:
Optional Image URL:
Upload some images for this post


[ Follow Ups ] [ Post Followup ] [ CPU-Central Message Board ]