Sempron 3300+ @2.7GHz

Zap

Elite Member
Oct 13, 1999
22,377
7
81
Sempron 3300+ E3 tray

Got that CPU from Monarch for $79+$1.99 shipping, so basically $81 shipped. Ordered on Thursday and was really suprised to see it Monday morning. My last order from Monarch showed up on Pony Express... about 1½ weeks or so (mobo/CPU combo).

Testing this chip in my Epox 8KDA3I board in 5MHz HTT increments and on the default detected vcore (which it claims to be 1.5v ???) it will POST at 2.8GHz, but not at 2.85GHz. Locked on Prime at 2.75GHz with the minimum 0.05v vcore boost, but at 2.70Ghz is passing Prime. With this CPU installed, the board gave me a bunch of new memory settings above DDR400, up to DDR500. Wish it had extra below it, but ahhh well.

Oh yeah, tested multipliers and it works so this has Cool and Quiet - we already knew that right?

Using the retail box HSF from my Sempron 2600+ and running Prime it's barely warm.

I think I can safely say that for new budget system builds, socket A is officially dead. Highest end is the Sempron 3000+ Barton core at 2GHz 333MHz FSB.

Pop one of these under $100 Semprons in a $60 Biostar board with a mild overclock and you've got the basis of a nice little system that'll be sufficient for most non-gamers.

I think paying a bit extra for a higher multiplier is worth the money. Lots of people are getting the 2800+ and then finding that they're hitting the motherboard limit before the CPU limit, so some go out and spend extra on a better overclocker board like the DFI. I say instead of spending the extra on the board, spend the extra on the CPU itself and get the next up multiplier such as the 3000+/3100+. Pretty much any board that's an overclocker board can hit 270-280MHz HTT for 2.4-2.5GHz clocks with the 9x multiplier. The better boards can hit around 300MHz for 2.7GHz which may be the limit of my single sample (not representative, but a possibility). If you get the 2600+/2800+ with the 8x multiplier you'd have to be running darn near 340MHz HTT to hit around 2.7GHz.

I'm not sure that paying extra for the 10x multiplier is worth it because at the limit of my particular chip the board is not my limiting factor. However, in my case it really wasn't paying extra since it was so inexpensive.

Too bad this chip didn't overclock to an even 3GHz. :brokenheart:

Next up on the test bench... my Biostar Tforce6100 and Sapphire Radeon X800GTO Ultimate Edition from Newegg refurb... board didn't come with ATX back plate but the video card looks brand new in box with everything.

Later this week, another Sapphire Radeon X800GTO (normal edition?) also from Newegg refurb... gosh there are a LOT of Radeon X8-- series on refurb for pretty cheap. What's up with that? We're talking about X800 128MB for $82, up to the X800XL with 512MB for $170 - I didn't even know they made 512MB versions of those...

Sometime next week, Monarch shipping department willing, my first Opteron, a 144 for $130 and a DFI LANPARTY Ultra-D for $120. My cheapest socket 939 CPU to date and my most expensive motherboard since the early socket 478 800MHz FSB days...

Good times, good times...
 

kmmatney

Diamond Member
Jun 19, 2000
4,363
1
81
Nice overclock! Is that a 64-bit processor? What HSF are you using?

I went through the same decisions, but decided on getting a "medium" priced MB and a Sempron 2800+. At the time, the 3300+ was a bit more expensive, and I didn't like the 128KB cache, and had no HSF. I run my Abit with the FSB at 300 MHz, HTT at 3x, and memory strarting at 133 MHz so the overclock brings it to 200 MHz. I'm only using the stock HSF so I stopped overclocking at 2.4 GHz.

One thing to look at is the effect of the cache size. If you look at the charts over at Tom's Hardware, you can gather enough data to predict how a 128KB cache Sempron will perform versus a 256MB Sempron. Before I bought my cpu, I gathered data for games, Video encoding, and multithreading, and the made projections on the overclocked performance. The trend lines work out very well. and here's what I got:

Gaming:

Far Cry (frames per second)

Sempron 128KB @ 1.6 = 134.4
Sempron 128KB @ 1.8 = 145.4
Sempron 128KB @ 2.0 = 155 (last real value - rest are projections)
Sempron 128KB @ 2.4 = 175.93
Sempron 128KB @ 2.5 = 181.08
Sempron 128KB @ 2.6 = 186.23
Sempron 128KB @ 2.7 = 191.38

Sempron 256KB @ 1.4 = 126.9
Sempron 256KB @ 1.6 = 140
Sempron 256KB @ 1.8 = 151.6
Sempron 256KB @ 2.0 = 162.7 (last real value - rest are projections)
Sempron 256KB @ 2.4 = 186.95
Sempron 256KB @ 2.5 = 192.9
Sempron 256KB @ 2.6 = 198.85
Sempron 256KB @ 2.7 = 204.8

Venice 3000+ = 164.6
Venice 3200+ = 176.5
Venice 3500+ = 186.2
San Diego 3700+ = 190.9


XVid Video Encoding (time in seconds)

time to complete, seconds - smaller is better

Sempron 128KB @ 1.6 = 510
Sempron 128KB @ 1.8 = 463
Sempron 128KB @ 2.0 = 425 (last real value)
Sempron 128KB @ 2.4 = 338.5
Sempron 128KB @ 2.5 = 317.25
Sempron 128KB @ 2.6 = 296
Sempron 128KB @ 2.7 = 274.75


Sempron 256KB @ 1.4 = 570
Sempron 256KB @ 1.6 = 508
Sempron 256KB @ 1.8 = 461
Sempron 256KB @ 2.0 = 425 (last real value - rest are projections)
Sempron 256KB @ 2.4 = 322.3
Sempron 256KB @ 2.5 = 298.2
Sempron 256KB @ 2.6 = 274.1
Sempron 256KB @ 2.7 = 250

Venice 3000+ = 437
Venice 3200+ = 398
Venice 3500+ = 366
San Diego 3700+ = 363


Multithreading - 4 threads

Time to complete, seconds - smaller is better

Sempron 128KB @ 2.4 = 729
Sempron 128KB @ 2.5 = 718
Sempron 128KB @ 2.6 = 707
Sempron 128KB @ 2.7 = 696

Sempron 256KB @ 2.4 = 529
Sempron 256KB @ 2.5 = 490.5
Sempron 256KB @ 2.6 = 452
Sempron 256KB @ 2.7 = 413.5

Venice 3000+ = 745
Venice 3200+ = 677
Venice 3500+ = 596
SanDiego 3700+ = 599 (? Tom's Hardware screw up?)



Multithreading - 2 threads

Time to complete, seconds - smaller is better

Sempron 128KB @ 2.4 = 177
Sempron 128KB @ 2.5 = 165.5
Sempron 128KB @ 2.6 = 154
Sempron 128KB @ 2.7 = 142.5

Sempron 256KB @ 2.4 = 160.9
Sempron 256KB @ 2.5 = 147.35
Sempron 256KB @ 2.6 = 133.8
Sempron 256KB @ 2.7 = 120.25

Venice 3000+ = 228
Venice 3200+ =208
Venice 3500+ =191
SanDiego 3700+ = 187


Summary:

* Games: A Sempron with 256KB of cache running at 2.5 GHz is about the same as a 128KB cached Sempron running at 2.7 GHz

* Encoding: A Sempron with 256KB running at 2.6 GHz is about the same as a 128KB cached Sempron running at 2.7 GHz

* MultiThreading:

- 2 Applications running: A Sempron with 256KB of cache needs to run at 2.54 GHz to match a 128 KB Sempron running at 2.7 GHz.

- 4 Applications running: A Sempron with 256KB of cache scales much better, witha 2.0 GHz 256KB Sempron matching a 128KB Sempron running at 2.7 GHz.



To Do: Predict scaling on Venice and Sand Diego cpus!
 

Avalon

Diamond Member
Jul 16, 2001
7,571
178
106
Very nice. My old Palermo and DFI NF3 250GB did 2.7, but that took some volts.
:)

If I'm going to get 2.7ghz either way, I'd rather pay more for the mobo than the CPU to do it, but that's just me. You did get CnQ while I didn't, though, which must be a nice feature.
 

3chordcharlie

Diamond Member
Mar 30, 2004
9,859
1
81
Originally posted by: kmmatney
Nice overclock! Is that a 64-bit processor? What HSF are you using?

I went through the same decisions, but decided on getting a "medium" priced MB and a Sempron 2800+. At the time, the 3300+ was a bit more expensive, and I didn't like the 128KB cache, and had no HSF. I run my Abit with the FSB at 300 MHz, HTT at 3x, and memory strarting at 133 MHz so the overclock brings it to 200 MHz. I'm only using the stock HSF so I stopped overclocking at 2.4 GHz.

One thing to look at is the effect of the cache size. If you look at the charts over at Tom's Hardware, you can gather enough data to predict how a 128KB cache Sempron will perform versus a 256MB Sempron. Before I bought my cpu, I gathered data for games, Video encoding, and multithreading, and the made projections on the overclocked performance. The trend lines work out very well. and here's what I got:

Gaming:

Far Cry (frames per second)

Sempron 128KB @ 1.6 = 134.4
Sempron 128KB @ 1.8 = 145.4
Sempron 128KB @ 2.0 = 155 (last real value - rest are projections)
Sempron 128KB @ 2.4 = 175.93
Sempron 128KB @ 2.5 = 181.08
Sempron 128KB @ 2.6 = 186.23
Sempron 128KB @ 2.7 = 191.38

Sempron 256KB @ 1.4 = 126.9
Sempron 256KB @ 1.6 = 140
Sempron 256KB @ 1.8 = 151.6
Sempron 256KB @ 2.0 = 162.7 (last real value - rest are projections)
Sempron 256KB @ 2.4 = 186.95
Sempron 256KB @ 2.5 = 192.9
Sempron 256KB @ 2.6 = 198.85
Sempron 256KB @ 2.7 = 204.8

Venice 3000+ = 164.6
Venice 3200+ = 176.5
Venice 3500+ = 186.2
San Diego 3700+ = 190.9


XVid Video Encoding (time in seconds)

time to complete, seconds - smaller is better

Sempron 128KB @ 1.6 = 510
Sempron 128KB @ 1.8 = 463
Sempron 128KB @ 2.0 = 425 (last real value)
Sempron 128KB @ 2.4 = 338.5
Sempron 128KB @ 2.5 = 317.25
Sempron 128KB @ 2.6 = 296
Sempron 128KB @ 2.7 = 274.75


Sempron 256KB @ 1.4 = 570
Sempron 256KB @ 1.6 = 508
Sempron 256KB @ 1.8 = 461
Sempron 256KB @ 2.0 = 425 (last real value - rest are projections)
Sempron 256KB @ 2.4 = 322.3
Sempron 256KB @ 2.5 = 298.2
Sempron 256KB @ 2.6 = 274.1
Sempron 256KB @ 2.7 = 250

Venice 3000+ = 437
Venice 3200+ = 398
Venice 3500+ = 366
San Diego 3700+ = 363


Multithreading - 4 threads

Time to complete, seconds - smaller is better

Sempron 128KB @ 2.4 = 729
Sempron 128KB @ 2.5 = 718
Sempron 128KB @ 2.6 = 707
Sempron 128KB @ 2.7 = 696

Sempron 256KB @ 2.4 = 529
Sempron 256KB @ 2.5 = 490.5
Sempron 256KB @ 2.6 = 452
Sempron 256KB @ 2.7 = 413.5

Venice 3000+ = 745
Venice 3200+ = 677
Venice 3500+ = 596
SanDiego 3700+ = 599 (? Tom's Hardware screw up?)



Multithreading - 2 threads

Time to complete, seconds - smaller is better

Sempron 128KB @ 2.4 = 177
Sempron 128KB @ 2.5 = 165.5
Sempron 128KB @ 2.6 = 154
Sempron 128KB @ 2.7 = 142.5

Sempron 256KB @ 2.4 = 160.9
Sempron 256KB @ 2.5 = 147.35
Sempron 256KB @ 2.6 = 133.8
Sempron 256KB @ 2.7 = 120.25

Venice 3000+ = 228
Venice 3200+ =208
Venice 3500+ =191
SanDiego 3700+ = 187


Summary:

* Games: A Sempron with 256KB of cache running at 2.5 GHz is about the same as a 128KB cached Sempron running at 2.7 GHz

* Encoding: A Sempron with 256KB running at 2.6 GHz is about the same as a 128KB cached Sempron running at 2.7 GHz

* MultiThreading:

- 2 Applications running: A Sempron with 256KB of cache needs to run at 2.54 GHz to match a 128 KB Sempron running at 2.7 GHz.

- 4 Applications running: A Sempron with 256KB of cache scales much better, witha 2.0 GHz 256KB Sempron matching a 128KB Sempron running at 2.7 GHz.



To Do: Predict scaling on Venice and Sand Diego cpus!
There's something wrong with your numbers.

At 'real' speeds, there seems to be no difference in encoding numbers between 128kb and 256kb semprons; then you project significant gains for the 256kb versions when overclocked.

Similarly, you have a huge projcted jump from 2 to 2.4 ghz for gaming with the 256kb version. Maybe there's a mistake in the numbers or they aren't 'lined up' properly or something.

At best I think the 1.4ghz data point is messy up your trend line; at worst there seems to be something wrong with the way you generated it.

I'm a big fan of more cache, and would definitely lean towards the 256kb models, (I still haven't seen numbers, but I expect a bigger performance hit for the smaller cahce under 64bit OS's), but I think your numbers may be a little optimistic.
 

Zap

Elite Member
Oct 13, 1999
22,377
7
81
I've got some benchmark numbers from the much reviled 3DMark05... this is actually from another Sempron I have, a 2600+ that also has 128k cache, overclocked on my Biostar Tforce6100 board that I mentioned and the Sapphire X800GTO Ultimate, also overclocked. That Ultimate video card needed additional cooling for the high overclock.

6060 3DMark05

Picture of setup

Screenshot of score & overclocks