PDA

View Full Version : AMD Athlon 64 3800+ Venice 90nm


W1zzard
05-15-2005, 04:08 PM
[page=Introduction]
Introduction
First, we could like to thank Mips-Computer (http://www.mips-computer.de) for supplying us with the CPU.

Advanced Micro Systems (AMD) has released a new revision of their Athlon64 S939, the code name is Venice. Venice is produced in 90nm, has 512KB Cache and is clocked betwen 1.8 GHz and 2.4 GHz. Compared to its predecessor, the Winchester, new features are support for SSE3 and an improved Integrated Memory Controller (IMC), which means the the CPU itself manages the computer's memory. With those improvements it is now possible to run 2GB memory at 200 MHz with a CommandRate of 2T. Previous CPUs were only able to run 2GB Memory in Async Mode (5:4 - Ram: 166 MHz 2T). We will test, if that has any effect on the memory's overclockability and if SSE3 does actually increase performance.

http://www.techpowerup.com/reviews/AMD/3800Venice/images/packpacked_small.jpg (http://www.techpowerup.com/reviews/AMD/3800Venice/images/packpacked_big.jpg) http://www.techpowerup.com/reviews/AMD/3800Venice/images/package_small.jpg (http://www.techpowerup.com/reviews/AMD/3800Venice/images/package_big.jpg) http://www.techpowerup.com/reviews/AMD/3800Venice/images/cpupack_small.jpg (http://www.techpowerup.com/reviews/AMD/3800Venice/images/cpupack_big.jpg) http://www.techpowerup.com/reviews/AMD/3800Venice/images/cpu_small.jpg (http://www.techpowerup.com/reviews/AMD/3800Venice/images/cpu_big.jpg) http://www.techpowerup.com/reviews/AMD/3800Venice/images/cpuback_small.jpg (http://www.techpowerup.com/reviews/AMD/3800Venice/images/cpuback_big.jpg)

<table border="1" cellpadding="3" cellspacing="0" class="ramtable">
<tr align="center">
<th colspan="2" scope="row" style="font-size:larger;text-align:center">Processor Details</th>
</tr>
<tr>
<th width="138" scope="row">Processor</th>
<td scope="row">AMD Athlon 64</td>
</tr>
<tr>
<th scope="row">Model</th>
<td scope="row">3800+</td>
</tr>
<tr>
<th scope="row">Ordering P/N (Tray)</th>
<td scope="row">ADA3800DAA4BP</td>
</tr>
<tr>
<th valign="top" scope="row">Ordering P/N (PIB)</th>
<td valign="top" scope="row">ADA3800BPBOX</td>
</tr>
<tr>
<th valign="top" scope="row">Rev</th>
<td valign="top" scope="row">E3 (Venice)</td>
</tr>
<tr>
<th valign="top" scope="row">Clock Frequency</th>
<td valign="top" scope="row">2400Mhz</td>
</tr>
<tr>
<th valign="top" scope="row">System Bus</th>
<td valign="top" scope="row">2000</td>
</tr>
<tr>
<th valign="top" scope="row">Voltage</th>
<td valign="top" scope="row">Variable</td>
</tr>
<tr>
<th valign="top" scope="row">Max Temp</th>
<td valign="top" scope="row">65°C</td>
</tr>
<tr>
<th valign="top" scope="row">Thermal Power</th>
<td valign="top" scope="row">89W</td>
</tr>
<tr>
<th valign="top" scope="row">L1 Cache</th>
<td valign="top" scope="row">128KB</td>
</tr>
<tr>
<th valign="top" scope="row">L2 Cache</th>
<td valign="top" scope="row">512KB</td>
</tr>
<tr>
<th valign="top" scope="row">CMOS Technology</th>
<td valign="top" scope="row">90nm SOI</td>
</tr>
<tr>
<th valign="top" scope="row">Package Profile</th>
<td valign="top" scope="row">Socket 939</td>
</tr>
</table>

Venice Model Chart
<table width="640" border="1" cellpadding="3" cellspacing="0" class="ramtable">
<tr>
<th nowrap="nowrap" scope="row" width="128">OPN (Tray)</th>
<th nowrap="nowrap" scope="row" width="108">Processor</th>
<th scope="row">Model</th>
<th scope="row">Clock Freq.</th>
<th scope="row" width="107">Package Profil</th>
<th scope="row" width="57">Voltage</th>
<th scope="row">Wattage</th>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="128"><span style="font-weight: 400">
ADA3800DAA4BP</span></th>
<td nowrap="nowrap" scope="row" width="108"><span style="font-weight: 400">
AMD Athlon 64</span></td>
<td scope="row">3800+</td>
<td scope="row">2400Mhz</td>
<td scope="row" width="107">Socket 939</td>
<td scope="row" width="57">Variable</td>
<td scope="row">89W</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row">
<span style="font-weight: 400">ADA3500DAA4BP</span></th>
<td nowrap="nowrap" scope="row"><span style="font-weight: 400">
AMD Athlon 64</span></td>
<td scope="row">3500+</td>
<td scope="row">2200Mhz</td>
<td scope="row">Socket 939</td>
<td scope="row">Variable</td>
<td scope="row">67W</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="128"><span style="font-weight: 400">
ADA3200DAA4BP</span></th>
<td nowrap="nowrap" scope="row" width="108"><span style="font-weight: 400">
AMD Athlon 64</span></td>
<td scope="row">3200+</td>
<td scope="row">2000Mhz</td>
<td scope="row" width="107">Socket 939</td>
<td scope="row" width="57">Variable</td>
<td scope="row">67W</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="128"><span style="font-weight: 400">
ADA3000DAA4BP</span></th>
<td nowrap="nowrap" scope="row" width="108"><span style="font-weight: 400">
AMD Athlon 64</span></td>
<td scope="row">3000+</td>
<td scope="row">1800Mhz</td>
<td scope="row" width="107">Socket 939</td>
<td scope="row" width="57">Variable</td>
<td scope="row">67W</td>
</tr>
</table>

Venice replaces current 512KB S939 CPUs Winchester (3000+, 3200+, 3500+) and NewCastle (3800+).

[page=Features & Test Systems]
Features
The Athlon64 Series offers a few nice features which we are going to discuss now.

SSE3
The Streaming SIMD Extensions 3 which AMD got in exchange for AMD64 (Intel: EMT64) from Intel is now used in the AMD CPUs. According to Intel multimedia and multi-threaded applications benefit most from SSE3.

NX-Bit - "Enhanced Virus Protection"
AMD introduced the NX-Bit to "improve computer security". This function works only when Windows XP with SP2 is installed and tries to stop execution of malicious code (viruses). Unfortunately this protection "works" only with viruses which use Buffer Overflow techniques.

Cool'n'Quiet
The advantage of Cool'n'Quiet is, that the CPU is clocked down (to 5x200 - 1000 Mhz at 1.1V), when no computation power is needed. This reduces heat output and makes it easier to keep the CPU cool, or decrease noise levels and power consumption. Once power is needed, the clock returns back to normal, basically on-demand performance.

AMD64
The AMD64 or x86-64 or x64 is a 64-bit processor architecture invented by AMD. It is a superset of the x86 architecture, which it natively supports. The AMD64 Instruction set is currently used in AMD's Athlon 64, Athlon 64 FX, and Opteron processors.

AMD64 represents a break with AMD's past behavior of following Intel's standards, but follows Intel's earlier behavior of extending the x86 architecture, from the 16-bit 8086 to the 32-bit 80386 and beyond, without ever removing backwards compatibility. The AMD64 architecture extends the 32-bit x86 architecture (IA-32) by adding 64-bit registers, with full 32-bit and 16-bit compatibility modes for earlier software. Source: Wikipedia


Test Systems

<table border="1" cellpadding="3" cellspacing="0" class="ramtable" width="450">
<tr align="center">
<th colspan="2" scope="row" style="font-size:larger;text-align:center">Test System "Venice"</th>
</tr>
<tr>
<th width="100" scope="row">CPU:</th>
<td scope="row">AMD Athlon64 3800+ (S939; 512KB; Venice)</td>
</tr>
<tr>
<th scope="row">Motherboard:</th>
<td scope="row">DFI Lanparty nForce4 SLI-D</td>
</tr>
<tr>
<th scope="row">Memory:</th>
<td scope="row">TwinMOS Twister PC3200 Pro (TCCD)</td>
</tr>
<tr>
<th valign="top" scope="row">Video Card:</th>
<td valign="top" scope="row">2x Pine XFX 6800 GT SLI @ 450/1200</td>
</tr>
<tr>
<th valign="top" scope="row">Harddisk:</th>
<td valign="top" scope="row">2x 74GB WD Raptor RAID 0</td>
</tr>
<tr>
<th valign="top" scope="row">Power Supply:</th>
<td valign="top" scope="row">OCZ ADJ 520W</td>
</tr>
<tr>
<th valign="top" scope="row">Software:</th>
<td valign="top" scope="row">Windows XP SP2, Forceware 76.44</td>
</tr>
</table>

<table border="1" cellpadding="3" cellspacing="0" class="ramtable" width="450">
<tr align="center">
<th colspan="2" scope="row" style="font-size:larger;text-align:center">Test System "Winchester"</th>
</tr>
<tr>
<th width="100" scope="row">CPU:</th>
<td scope="row">AMD Athlon64 3200+ (S939; 512KB; Winchester)</td>
</tr>
<tr>
<th scope="row">Motherboard:</th>
<td scope="row">DFI Lanparty nForce4 SLI-D</td>
</tr>
<tr>
<th scope="row">Memory:</th>
<td scope="row">TwinMOS Twister PC3200 Pro (TCCD)</td>
</tr>
<tr>
<th valign="top" scope="row">Video Card:</th>
<td valign="top" scope="row">2x Pine XFX 6800 GT SLI @ 450/1200</td>
</tr>
<tr>
<th valign="top" scope="row">Harddisk:</th>
<td valign="top" scope="row">2x 74GB WD Raptor RAID 0</td>
</tr>
<tr>
<th valign="top" scope="row">Power Supply:</th>
<td valign="top" scope="row">OCZ ADJ 520W</td>
</tr>
<tr>
<th valign="top" scope="row">Software:</th>
<td valign="top" scope="row">Windows XP SP2, Forceware 76.44</td>
</tr>
</table>
Our AMD Test Systems are fully watercooled (30°C water). For CPU cooling a HV3 waterblock from 1A-Cooling is used. Modern air coolers with a high-flow fan are almost comparable to watercooling.

<table border="1" cellpadding="3" cellspacing="0" class="ramtable" width="450">
<tr align="center">
<th colspan="2" scope="row" style="font-size:larger;text-align:center">Test System "P4 2.4C"</th>
</tr>
<tr>
<th width="100" scope="row">CPU:</th>
<td scope="row">Intel Pentium 4 2.4C (S478; 512KB; Northwood)</td>
</tr>
<tr>
<th scope="row">Motherboard:</th>
<td scope="row">ABIT IC7, Bios 2.8</td>
</tr>
<tr>
<th scope="row">Memory:</th>
<td scope="row">2x 256MB Generic PC3200 2.5-3-3-6</td>
</tr>
<tr>
<th valign="top" scope="row">Video Card:</th>
<td valign="top" scope="row">ATI X800 XT</td>
</tr>
<tr>
<th valign="top" scope="row">Harddisk:</th>
<td valign="top" scope="row">Maxtor DiamondMax 160GB</td>
</tr>
<tr>
<th valign="top" scope="row">Power Supply:</th>
<td valign="top" scope="row">Antec TrueControl 550W</td>
</tr>
<tr>
<th valign="top" scope="row">Software:</th>
<td valign="top" scope="row">Windows XP SP2, Catalyst 5.4</td>
</tr>
<tr>
<td colspan="2">OC 3400 is 15x283 FSB, Mem Ratio 2:3 (=188 MHz)</td>
</tr>
</table>

<table border="1" cellpadding="3" cellspacing="0" class="ramtable" width="450">
<tr align="center">
<th colspan="2" scope="row" style="font-size:larger;text-align:center">Test System "AA8XE"</th>
</tr>
<tr>
<th width="100" scope="row">CPU:</th>
<td scope="row">Intel Pentium 4 3.0F (S775; 2MB; Prescott)</td>
</tr>
<tr>
<th scope="row">Motherboard:</th>
<td scope="row">ABIT Fatal1ty AA8XE, Bios 1.4</td>
</tr>
<tr>
<th scope="row">Memory:</th>
<td scope="row">2x 512MB OCZ PC2-5400 EB 4-2-2-8</td>
</tr>
<tr>
<th valign="top" scope="row">Video Card:</th>
<td valign="top" scope="row">ATI X800 Non-Pro PCI-E</td>
</tr>
<tr>
<th valign="top" scope="row">Harddisk:</th>
<td valign="top" scope="row">Maxtor DiamondMax 160GB</td>
</tr>
<tr>
<th valign="top" scope="row">Power Supply:</th>
<td valign="top" scope="row">HEC Power475</td>
</tr>
<tr>
<th valign="top" scope="row">Software:</th>
<td valign="top" scope="row">Windows XP SP2, Catalyst 5.4</td>
</tr>
<tr>
<td colspan="2">OC 4500 is 15x300 FSB, Mem Ratio 1:1 (=300 MHz)</td>
</tr>
</table>

[page=Memory Controller]
Memory Controller
Here we compare the Venice's IMC with the one integrated into the Winchester. Used memory is TwinMOS Twister Pro with TCCD Chips. Samsung's TCCD Chips are known for their high clock rates and that they are able to run tight 2-2-2-5-7-256-1T timings at 200 MHz.
We reduced the multiplier on our 3800+, because we only had a 3200+ Winchester to compare to.

<table width="640" border="1" cellpadding="3" cellspacing="0" class="ramtable">
<tr>
<th nowrap="nowrap" scope="row">CPU Clock &amp; <br />
Memory Ratio</th>
<th scope="row">Memory<br />
Speed</th>
<th scope="row">Memory<br />
Timings</th>
<th scope="row">Everest<br />
Read</th>
<th scope="row">Everest<br />
Write</th>
<th scope="row">Everest<br />
Latency</th>
</tr>
<tr>
<th nowrap="nowrap" scope="row">Winchester (3200+) 10 x 200 1:1</th>
<td scope="row">200 MHz</td>
<td scope="row">2-2-2-5 2.8V</td>
<td scope="row">5578 MB/s</td>
<td scope="row">2340 MB/s</td>
<td scope="row">47.2 ns</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row">Venice (3200+) 10 x 200 1:1</th>
<td scope="row">200 MHz</td>
<td scope="row">2-2-2-5 2.8V</td>
<td scope="row">5369 MB/s</td>
<td scope="row">2393 MB/s</td>
<td scope="row">47.3 ns</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row">Winchester (3000+) 9 x 200 1:1</th>
<td scope="row">200 MHz</td>
<td scope="row">2-2-2-5 2.8V</td>
<td scope="row">5419 MB/s</td>
<td scope="row">2290 MB/s</td>
<td scope="row">48.8 ns</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row">Venice (3000+) 9 x 200 1:1</th>
<td scope="row">200 MHz</td>
<td scope="row">2-2-2-5 2.8V</td>
<td scope="row">5126 MB/s</td>
<td scope="row">2306 MB/s</td>
<td scope="row">49.3 ns</td>
</tr>
</table>

http://www.techpowerup.com/reviews/AMD/3800Venice/images/everestread.gif http://www.techpowerup.com/reviews/AMD/3800Venice/images/everestwrite.gif http://www.techpowerup.com/reviews/AMD/3800Venice/images/everestlatency.gif

The Winchester has a higher read bandwith. Actually we expected a different picture, because we thought that the Venice has an improved Memory Controller. This does not seem to be the case.

Now we tried to find out what the maximum clock speeds are at relaxed 2.5-4-4-5-7-256-1T timings with 2.8V Vdimm. To make sure that the CPU's overclocking limits to do not hinder our test, we set the multiplier to 8 and Vcore to 1.5v.

<table width="640" border="1" cellpadding="3" cellspacing="0" class="ramtable">
<tr>
<th nowrap="nowrap" scope="row">CPU Clock &amp; <br />
Memory Ratio</th>
<th scope="row">Memory<br />
Speed</th>
<th scope="row">Memory<br />
Timings</th>
<th scope="row">Everest<br />
Read</th>
<th scope="row">Everest<br />
Write</th>
<th scope="row">Everest<br />
Latency</th>
</tr>
<tr>
<th nowrap="nowrap" scope="row">Winchester 8 x 305 1:1</th>
<td scope="row">305 MHz</td>
<td scope="row">2,5-4-4-5 2.8V</td>
<td scope="row">7041 MB/s</td>
<td scope="row">3225 MB/s</td>
<td scope="row">43.0 ns</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row">Venice 8 x 308 1:1</th>
<td scope="row">308 MHz</td>
<td scope="row">2,5-4-4-5 2.8V</td>
<td scope="row">6622 MB/s</td>
<td scope="row">3237 MB/s</td>
<td scope="row">42.6 ns</td>
</tr>
</table>
http://www.techpowerup.com/reviews/AMD/3800Venice/images/maxmem.gif http://www.techpowerup.com/reviews/AMD/3800Venice/images/max_everestread.gif http://www.techpowerup.com/reviews/AMD/3800Venice/images/max_everestwrite.gif http://www.techpowerup.com/reviews/AMD/3800Venice/images/max_everestlatency.gif

You can see a small advantage of 3 MHz. We do not know, if this is random luck or caused by the Venice core.

We saw no issues running four 512 MB modules (double sided) at 200 MHz with timings of 2.5-2-2-7 (Crucial Ballistix Tracer PC4000), but on both CPUs only with a CommandRate of 2T.

We could not find any advantage of the "new" memory controller in the Venice. That's a bit dissapointing. AMD could have certainly improved things, for example run 4x512MB at 1T.

[page=Overclocking]
Overclocking
AMD's Athlon64 CPUs are known for their overclockability. We want to find out now, if that is the case with the Venice as well.
We reduced the multiplier to 11 for realistic results, since most overclockers use DDR500 and faster, a higher memory clock increases performance as well.

2400 MHz is the default clock speed. With settings at default, the board supplies 1.36V VCore to the CPU. Unfortunately it is unclear what the default voltage for the processor is, because AMD (http://www.amdcompare.com/us-en/desktop/default.aspx) lists it as "variable".

http://www.techpowerup.com/reviews/AMD/3800Venice/images/overclocking.gif

<table width="400" border="1" cellpadding="3" cellspacing="0" class="ramtable">
<tr>
<th nowrap="nowrap" scope="row" width="136">Clock Freq.</th>
<th nowrap="nowrap" scope="row" width="91">CPU Clock &amp; <br />
Memory Ratio</th>
<th scope="row" width="52">Voltage</th>
<th scope="row" width="59">Temperature</th>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="136">2400Mhz (default)</th>
<td nowrap="nowrap" scope="row" width="91"><span style="font-weight: 400">
12 x 200 1:1</span></td>
<td scope="row" width="52">1.36v</td>
<td scope="row" width="59">44°C</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="136">2436Mhz</th>
<td nowrap="nowrap" scope="row" width="91">
<span style="font-weight: 400">12 x 203 1:1</span></td>
<td scope="row" width="52">1.20v</td>
<td scope="row" width="59">38°C</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="136">2684Mhz</th>
<td nowrap="nowrap" scope="row" width="91"><span style="font-weight: 400">
11 x 244 1:1</span></td>
<td scope="row" width="52">1.36v </td>
<td scope="row" width="59">46°C</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="136">2750Mhz</th>
<td nowrap="nowrap" scope="row" width="91"><span style="font-weight: 400">
11 x 250 1:1</span></td>
<td scope="row" width="52">1.40v</td>
<td scope="row" width="59">47°C</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="136">2794Mhz</th>
<td nowrap="nowrap" scope="row" width="91"><span style="font-weight: 400">
11 x 254 1:1</span></td>
<td scope="row" width="52">1.45v</td>
<td scope="row" width="59">49°C</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="136">2816Mhz</th>
<td nowrap="nowrap" scope="row" width="91"><span style="font-weight: 400">
11 x 256 1:1</span></td>
<td scope="row" width="52">1.50v</td>
<td scope="row" width="59">51°C</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="136">2838Mhz</th>
<td nowrap="nowrap" scope="row" width="91"><span style="font-weight: 400">
11 x 258 1:1</span></td>
<td scope="row" width="52">1.54</td>
<td scope="row" width="59">54°C</td>
</tr>
<tr>
<th nowrap="nowrap" scope="row" width="136">2860Mhz</th>
<td nowrap="nowrap" scope="row" width="91"><span style="font-weight: 400">
11 x 260 1:1</span></td>
<td scope="row" width="52">1.57</td>
<td scope="row" width="59">56°C</td>
</tr>
</table>


http://www.techpowerup.com/reviews/AMD/3800Venice/images/suerpi2860_small.jpg (http://www.techpowerup.com/reviews/AMD/3800Venice/images/suerpi2860_big.jpg)
2860Mhz; 286x10; 1:1; 2,5-3-3-5-1T

Very good results. Even at 1.2V, which is almost 0.2V below standard voltage, it was possible to overclock the processor a little bit. This is important for people who want a silent system.
When overclocking, we could increase the clocks by almost 20%, and this on AMD's top model.

During overclocking the Venice got pretty warm. That's why we reduced water temperature to 25°C (in a normal water cooling system, it is not possible to reach this temperature).
This made a difference. Now, with a voltage of 1.62V, we were able to run stable at 2900 MHz, SuperPi even ran fine at 2950 Mhz. The CPU reached a temperature of 53°C. It might be possible that if the heatspreader is removed, temperatures drop enough to be able to get these results with regular 30°C water cooling.
The big temperature difference between water and CPU suggests that the contact between IHS and CPU core is not ideal, so the CPU temperature rises because heat can not be removed fast enough.

http://www.techpowerup.com/reviews/AMD/3800Venice/images/290x10_162_small.jpg (http://www.techpowerup.com/reviews/AMD/3800Venice/images/290x10_162_big.jpg) http://www.techpowerup.com/reviews/AMD/3800Venice/images/superpi2950_small.jpg (http://www.techpowerup.com/reviews/AMD/3800Venice/images/superpi2950_big.jpg)
2900Mhz; 290x10; 1:1; 2.5-3-3-5-1T 2950Mhz; 295x10; 1:1; 2.5-3-3-5-1T
[page=Performance: Sandra]
SiSoftware Sandra
http://www.techpowerup.com/reviews/AMD/3800Venice/images/sandra1.gif http://www.techpowerup.com/reviews/AMD/3800Venice/images/sandra2.gif

At the same clocks, Venice and Winchester show the same performance.
[page=Performance: SuperPi]
SuperPi
http://www.techpowerup.com/reviews/AMD/3800Venice/images/superpi1m.gif http://www.techpowerup.com/reviews/AMD/3800Venice/images/superpi32m.gif

Venice is definitely faster than Winchester, this is the biggest difference we could see in all benchmarks. The Pentium 4s are slower.
[page=Performance: PCMark04 & 3DMark01]
PCMark 2004
http://www.techpowerup.com/reviews/AMD/3800Venice/images/pcmark.gif

In PCMark 2004 the Venice is only a tick faster than the Winchester, but both get beaten by the Pentium 4 CPUs which have HyperThreading.
3DMark 2001
http://www.techpowerup.com/reviews/AMD/3800Venice/images/3dmark01.gif

We see no difference between Venice and Winchester in 3DMark2001SE, but you can see how much difference a higher CPU clock makes here and that the P4 does not stand a chance against the Athlon64.
[page=Performance: Cinebench & KribiBench]
CineBench
http://www.techpowerup.com/reviews/AMD/3800Venice/images/cinebench.gif

Both CPU cores show no performance difference. Pentium 4 wins by HyperThreading.
KribiBench
http://www.techpowerup.com/reviews/AMD/3800Venice/images/kribibench.gif

[page=Performance: Comanche4 & Quake3]
Comanche 4
http://www.techpowerup.com/reviews/AMD/3800Venice/images/comanche4.gif

The Athlon64 can show its advantage in Comanche 4. Venice and Winchester are almost the same speed.
Quake 3 Arena
http://www.techpowerup.com/reviews/AMD/3800Venice/images/q3a.gif

Basically the same image as in Comanche 4. Athlon64 dominates, there is only a small difference between Venice and Winchester.
[page=Summary & Conclusion]
Summary
Venice shows not much of a performance gain in all benchmarks. Who already has a Winchester can keep it and wait till prices drop. At the moment, there are almost no applications or games which really use SSE3. Compared to the Pentium4 the Athlon64 is strong and shows the known picture - faster in games, weaker in office applications.

Conclusion
<table width="100%" cellpadding="5" cellspacing="0" id="result">
<tr><th>http://www.techpowerup.com/images/dollar.gif</th>
<td>
The processor is sold for $360 to $400 which is a pretty big premium, compared to the smaller models (3000+, 3200+).</td>
</tr><tr>
<th>http://www.techpowerup.com/images/thumbup.gif</th>
<td>

Overclockability
Performance
SSE3
</td>
</tr>
<tr>
<th>http://www.techpowerup.com/images/thumbdown.gif</th>
<td>
Price
Heat transfer between CPU and cooler (Heatspreader)
Memory Controller not improved
</td></tr>
<tr><th>8.3</th>
<td>The AMD Athlon64 3800+ with Venice core is geared towards performance hungry people, mostly gamers, which have the money. Overclockers will love the Venice. SSE3 makes a small difference, but the difference is so small, that upgrading from Wincehster to Venice is not worth it at the moment. People who are looking into getting a new Socket939 CPU should definitely consider the Venice.</td></tr>
<tr><th></th><td></td></tr>
</table>

gR3iF
05-15-2005, 04:50 PM
hmpf and i thought my oc result is good :D
but n1 cpu for future but at least no option to the x2 cpus and the coming things

Unregistered
05-17-2005, 11:37 AM
"an improved Integrated Memory Controller (IMC), which means the the CPU itself manages the computer's memory. With those improvements it is now possible to run 2GB memory at 200 MHz with a CommandRate of 2T. Previous CPUs were only able to run 2GB Memory in Async Mode (5:4 - Ram: 166 MHz 2T). "

????
I am running 4x512 Corsair XMS PC-550, e.g. 2GB Memory with a Winchester 3500+ on a MSI K8N
Platinum at 210MHz with 2T!!!!

how is this Possible?

greetz
ODO

nightelf84
05-17-2005, 12:56 PM
"an improved Integrated Memory Controller (IMC), which means the the CPU itself manages the computer's memory. With those improvements it is now possible to run 2GB memory at 200 MHz with a CommandRate of 2T. Previous CPUs were only able to run 2GB Memory in Async Mode (5:4 - Ram: 166 MHz 2T). "

????
I am running 4x512 Corsair XMS PC-550, e.g. 2GB Memory with a Winchester 3500+ on a MSI K8N
Platinum at 210MHz with 2T!!!!

how is this Possible?

greetz
ODO

And what kind of speed is ur memory running at in this configuration?

Unregistered
05-17-2005, 01:29 PM
CPU and RAM @ 210......

i just checked it, cause i was a littel bit confused!

greetz
ODO

Unregistered
05-17-2005, 01:46 PM
sorry, or do you mean the Timmings?

2,5-4-4-8

greetz
ODO

ZathaN
05-20-2005, 08:22 PM
> Venice is definitely faster than Winchester, this is the biggest difference we could see in all benchmarks. The Pentium 4s are slower. <

Intel Dothan is faster than any AMD64

gR3iF
05-20-2005, 08:53 PM
so what? tell us somethig new
like a good pentium m board

wazzledoozle
05-20-2005, 10:44 PM
Free Athlon 64 :eek:

Morlak
05-26-2005, 01:00 AM
> Venice is definitely faster than Winchester, this is the biggest difference we could see in all benchmarks. The Pentium 4s are slower. <

Intel Dothan is faster than any AMD64

Not quite faster but holding their own. Reviews have shown the M to be a very good processor, but the new AMD's still kick its little butt in almost every test.

I compared my own overclocked P4 650 (3.4Ghz) with the performance of the Venice 3800+ in the review here.

Processors cost about the same but I think the 3800+ is just a little bit cheaper.

Sandra CPU Benchmark

3800+ 13228 MIPS Highest Overclock 11x260
3800+ 12717 MPIS Lowest Overclock 11x244
Mine: 11873 MIPS
3800+: 11143 MIPS Stock 11x200

Mine: 4940/8324 Flops
3800+ 4517/5846 Flops Highest Overclock 11x260


SuperPI 1M

3800+ 29.7 Highest Overclock 11x260
3800+ 31.6 Lowest Overclock 11x244
Mine 33
3800+ 35.4 Stock

Super PI 32M

3800+ 28.98 Highest Overclock 11x260
3800+ 30.71 Stock 11x200
Mine 31.71
3500+ 33.06 11x200

Number crunching performance is not quite so lopsided as the review shows. You cannot say "P4s are slow", at least any recent P4 compared to this AMD. FX-55s and X2s oh yeah you bet. Everything is slow compared to them. But there is a huge difference in the 600 series compared to what is used here. I'll even try the benchmarks again when I don't have other stuff running. BeyondTV was recording a show, and my raid array was rebuilding a mirror, and all of that surely affected my results somewhat. Haha it's all in fun anyway. Id like to see how a overclocked 3.6 or 3.8Ghz P4 compares with a 3800+.

CTRL-ALT-DEL
07-12-2005, 09:46 PM
Is those temperatures under load or idle? If under load, what program did you guys use?

wazzledoozle
07-12-2005, 11:18 PM
You cant expect good temperatures with the stock cooler and thermal paste... (at 2.9 GHz)

Quake2owns
02-05-2006, 09:00 AM
You cant expect good temperatures with the stock cooler and thermal paste... (at 2.9 GHz)
xp-90c 92mm fan cpu temps 2.9 ghz=65c idle.