A lamp is workable but choose it carefully so it can itself do without cooling air. The heating will be slow.
A useful technique for quickly heating your device and avoiding the thermal inertia of a block of metal would be to use a 0.001 to 0.01" thick strip of perhaps copper, or another metal of your own choice that won't reduce or melt with heat, with a cutout for your laser diode's flanges to be mounted to, perhaps with a thin ring clamp top and bottom and very small screws. I say clamp because a solder with a high enough melting point may be too hot the the diode's internal construction materials (as pointed out above). I faced this sort of thing working with down-hole companies and that is one reason they often prefer to buy dice and do their own assembly. A high resistance metal is better because it will take more voltage and less current. Even so, 100-400 Amps of DC current is called for.
Anyway, pass a large current through the copper strip to heat it. A blower could cool it fairly quickly if needed. The thinner the material, the less current required for a given temperature rise. The thicker the material, the more uniform the temperature in the region around the diode. Consider that the case temperature of the diode might not be completely uniform due to radiation and convection.
To make the current, consider a large transformer, 1-3KVA. It is possible to wind 1-2 turns of bus bar and there you have more amps than you need. rectify, filter.. Or use an old fashioned carbon pile and a 2V industrial wet cell such as an old telco battery, or a 6V automotive battery.
Best practice would be to avoid passing current through the metal case/body of the diode. Meaning insulating it electrically and not thermally. At those temperatures it will require a little more work. perhaps a BeO or alumina heat sink insulator, but they are difficult to machine (BeO dust is very hazardous) and not usually found in the right size, most commonly for TO-220 and TO-3, and some newer semiconductor cases. Anyway, find a material for insulating transistor cases from heatsinks yet encourages thermal transfer, that will stand your temperatures. Even so, it might not harm anything to allow some incidental current to pass through the mounting flange as long as the current loops are kept separate.
Thin clamping and tiny hardware to reduce thermal inertia and uneven heating. While you are at it, a slight deformity or a tab in the clamp could hold the thermistor or whatever you are using to measure the laser body temperature.
Regulate the current to help regulate the temperature. Thin metal strips may deform and move about as heated so you may wish to mount the laser itself with ceramic or something that won't suck the heat away but will hold the laser in place well, and use flexible leads such as a fanned out welding cable or even a couple of braids from RG-8 cable to deliver the current to the metal strip.
There are several manufacturers of refractory ceramics that could supply the mounting material, some of which can even be drilled carefully. I assume no responsibility for destructive play.
I was at a solar technology trade show and stopped at the ZRCI booth. Their array of materials and possible shapes was impressive. I was kindly given a piece of silica-based RSLE-57 good to 1100 dec C that I want to try as a beam stop. Its pretty dense but some of the ceramic fiber based materials they supply might be better for your project. Some of the materials are machinable. zrci.com
), "these (wirewound) resistors can withstand temperatures of up to +450 °C" 
