Everest 5.00 released!

[Fiery]

I love this program. However, where was my email notification of this release!?!? I'm glad they fixed the latency issues that plagued the earlier betas.

The newsletter was sent out yesterday. I'm sorry for the delays, we had some issues with the mass emailing system we use

[eva2000]

Benchmark results should only be compared with results obtained with the same EVEREST version. EVEREST itself warns you about that each time you run a benchmark.

any way to disable that message as it gets annoying after the 1000000th time

[Fiery]

any way to disable that message as it gets annoying after the 1000000th time

And still, quite a few users don't recognize it Or maybe they don't interpret it

[zlojack]

I notice the latency is now reported much higher for i7 RAM.

[zanzabar]

And still, quite a few users don't recognize it Or maybe they don't interpret it

i like it still, its never more than 500MB/s or so off from another version

[3NZ0]

Cheers for the heads up, seems my old 4.6 key works with this as well.

[Fiery]

I notice the latency is now reported much higher for i7 RAM.

In previous EVEREST versions the special IMC optimizations of Core i7 processors managed to get around our memory latency measurement routine and optimized it beyond the expected level. That's why EVEREST v5.00 measures considerably higher memory latency scores on Core i7, and actually those high scores are the real ones. We expect the real memory latency to be at least 50% higher than the ones your could measure using EVEREST v4.60.

[Roger_D25]

Thanks for the great work Fiery, I'm still very happy to have purchased it! I could be just imagining things but v5.00 seems to be measuring lower latency on my system as well (with same exact settings as when I was using v4.60), could this be possible? It went from an average of 55.6ns down to 52.2ns (averaged three different runs, each run after a restart)? Although I've never measured the difference before this could simply be the difference that any new revision has, hence why you mentioned you should only compare data within the same revision version?

[Fiery]

We keep on improving our benchmark routines, so at a version update it is normal to have some benchmark methods producing better scores than with the previous EVEREST version.

[saaya]

very nice work on the i7 latency numbers, makes sense that its much higher actually looking at the performance of latency favouring applications

[felix_w]

Any plan on updating the way that the cores temperature is calculated on AMD cpu's ? Still shows numbers close to 10 °C or less.

I mean is there an offset to be dealt in the sensor reading (like Coretemp was updated recently ) ?

[Sunfire]

Any plan on updating the way that the cores temperature is calculated on AMD cpu's ? Still shows numbers close to 10 °C or less.

I mean is there an offset to be dealt in the sensor reading (like Coretemp was updated recently ) ?

Try this one

[eva2000]

In previous EVEREST versions the special IMC optimizations of Core i7 processors managed to get around our memory latency measurement routine and optimized it beyond the expected level. That's why EVEREST v5.00 measures considerably higher memory latency scores on Core i7, and actually those high scores are the real ones. We expect the real memory latency to be at least 50% higher than the ones your could measure using EVEREST v4.60.

indeed!

Super Pi 32M / Everest 4.60.1635 Beta



Super Pi 32M / Everest 5.00.1652 Beta