What gear do you use when tuning on the dyno?

I have a car with 6 gears. Started off tuning in 4th gear to 140kph. My dyno showed load at 55% and i am making just over 600NM of torque. However as soon as i do runs in 5th or 6th gear I make more power. In some cases torque is as high as 781NM. Howver areas that had no knock spikes they start to show knock spikes on data logger. Which gear do you guys use and how much load should i really put on the dyno so that when i go on the road i see same results as lately i been seeing knock spikes on the road but on dyno all is good

Normally you would use the gear with 1:1 gear ratio which provides direct correlation.

Agreed, 1:1 gear is the one.

Accuracy. A 1:1 gear ratio will give you the most accurate reading of your engine's power output. This is because there is no additional gear multiplication involved, so the dyno can accurately measure the power that is being produced by the engine.

Consistency. Using a 1:1 gear ratio will also help to ensure that your dyno runs are consistent. This is because the same gear ratio will be used for each run, so there will be no variation in the power output due to different gear ratios.

Safety. Using a 1:1 gear ratio can help to reduce the risk of accidents on the dyno. This is because the lower torque multiplication will make it less likely that the car will lose traction and spin out.

The 1:1 (direct) gear will give the least gearbox power loss, but that really isn't what you should be worried about.

Because you're looking at using the dyno' as a tuning tool, the highest gear that doesn't over-speed the dyno' will give the longest run and load the engine the most, and that's why I'd suggest that.

Then, if needed, you can fine tune things like the timing and fuelling on the road where you can load the engine fully.

While 1:1 has been the golden rule, it's not how I do things. I find it's beneficial to be flexible.

Some dyno cells have better cooling than others, some dynos apply more load than others, some dyno/vehicle combos may need a higher gear either for traction or to avoid maxing out the torque the dyno can handle.

I choose a gear that provides sufficient load, a pull of sufficient duration to gather good data, but without the dyno run taking too long so unnecessary abuse is avoided, and I avoid exceeding safe dyno speed or drivetrain/tire speed.

On some vehicles that's near 1:1, on most low/medium HP setups it's higher than 1:1 (shorter gear) to avoid the pull being abusively long.

I note which gear I used for each run so I can reference it at a later date, along with gear speed info if relevant for the dyno type.

On many vehicles I'll do pulls in a shorter gear til the tune is dialed in to reduce heat, then test in a longer gear, then run through multiple gears.